Description:FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:
A METHOD AND A SYSTEM FOR TISSUE CLASSIFICATION IN MEDICAL IMAGES
Applicant:
QURE.AI TECHNOLOGIES PRIVATE LIMITED
An Indian entity having address as:
6th Floor, 606, Wing E, Times Square, Andheri-Kurla Road, Marol, Andheri (E), Marol Naka, Mumbai, Mumbai, Maharashtra, India, 400059
The following specification particularly describes the invention and the manner in which it is to be performed. 
CROSS-REFERENCE TO RELATED APPLICATIONS AND PRIORITY
[0001] The present application does not claim priority from any other patent application.
FIELD OF INVENTION
[001] The present invention, in general, relates to the field of medical image processing and more particularly, relates to a method and a system for tissue classification in one or more medical images.

BACKGROUND OF THE INVENTION
[0002] This section is intended to introduce the reader to various aspects of art, which may be related to various aspects of the present disclosure that are described or claimed below. This discussion is believed to be helpful in providing the reader with background information to facilitate a better understanding of the various aspects of the present disclosure. Accordingly, it should be understood that these statements in this background section are to be read in this light, and not as admissions of prior art. Similarly, a problem mentioned in the background section or associated with the subject matter of the background section should not be assumed to have been previously recognized in the prior art. The subject matter in the background section merely represents different approaches, which in and of themselves may also correspond to implementations of the claimed technology.
[0003] Conventional imaging-based stroke diagnostic systems rely primarily on advanced imaging modalities such as Computed Tomography Perfusion (CTP) and Magnetic Resonance Imaging (MRI) with Diffusion-Weighted Imaging (DWI) and Perfusion-Weighted Imaging (PWI) to identify ischemic core and penumbra in acute stroke patients. While these imaging techniques provide highly accurate information, they are often inaccessible in resource-limited or rural healthcare settings. These modalities depend on availability of expensive, specialized equipment, require trained personnel, and often involve the use of contrast agents, which may not be suitable for all patients. As a result, access to timely and effective stroke diagnosis and treatment is limited for many patients, particularly those in smaller hospitals or low-resource environments.
[0004] Furthermore, the CT Perfusion imaging necessitates intravenous contrast dye, which can lead to adverse side effects. While most reactions are mild such as a metallic taste, warm flush, or transient urgency to urinate some patients may experience allergic responses or kidney complications. Individuals with pre-existing kidney disease or at risk of renal impairment, such as diabetic patients, face heightened danger. In rare cases, contrast-induced nephropathy may occur. These risks, along with the need for contrast agents, make CTP unsuitable for many vulnerable stroke patients. In addition, the time required for contrast injection and imaging interpretation significantly delays the initiation of stroke treatment in time-critical scenarios.
[0005] In the context of acute ischemic stroke, two critical zones of brain tissue are evaluated for clinical decision-making: the ischemic core and the penumbra. The ischemic core represents brain tissue that has already suffered irreversible damage due to insufficient blood flow. In contrast, the penumbra is a surrounding area of hypo perfused but still viable tissue that remains at risk of progressing to infarction if blood flow is not rapidly restored. The penumbra is considered salvageable if reperfusion therapies are administered promptly, making its rapid and accurate identification essential for effective intervention. Delays in identifying and treating the penumbral tissue can result in its irreversible conversion into core infarct, significantly worsening patient outcomes. Thus, time is of the essence in acute stroke care, and precise detection of penumbra can be the difference between recovery and permanent disability or death.
[0006] Similarly, MRI with DWI and PWI offers high-resolution insights into ischemic tissue but is limited by longer acquisition times, higher costs, and reduced availability in emergency contexts. The infrastructure required for MRI imaging is often unavailable in remote or under-equipped healthcare facilities, and the imaging process itself is less suited to rapid decision-making compared to CT-based alternatives. Moreover, MRI procedures require specially trained technicians and patient compliance, making it a less practical choice during acute stroke events where time is of the essence.
[0007] In clinical practice, non-contrast CT (NCCT) is the most widely used modality due to its speed and broad availability. However, the NCCT lacks inherent ability to visualize the ischemic core and the penumbra in early stages of stroke. Even experienced neuroradiologists are unable to identify subtle ischemic changes or assess tissue viability using the NCCT alone. This leads to over-reliance on clinical judgment, which can result in inconsistent diagnoses and delayed treatment. The absence of visible markers of the ischemic core and penumbra on the NCCT, hinders the timely identification of patients eligible for reperfusion therapies such as thrombolysis or mechanical thrombectomy.
[0008] While combining the NCCT with the CTP or the MRI can provide comprehensive insights, this approach retains the limitations of requiring contrast agents, specialized equipment, longer imaging times, and highly trained personnel. Moreover, it prevents the standalone use of the NCCT for advanced stroke diagnostics, thus failing to democratize access to timely and effective stroke care. The reliance on complex imaging workflows adds significant delays in the acute stroke pathway, with potentially irreversible consequences for patient outcomes.
[0009] The current stroke diagnosis paradigm lacks a scalable, accessible, and rapid solution for identifying the ischemic core and the penumbra using only NCCT images. There exists a critical unmet need for technology that can transform standard, widely available NCCT into a powerful tool for estimating ischemic regions, without the need for contrast injections or specialized hardware. Bridging this diagnostic gap would allow clinicians to make quicker and more accurate treatment decisions, particularly in resource-constrained settings where advanced imaging is not feasible.
[0010] In light of the above stated discussion, there exists a need for an improved system and a method for tissue classification in the one or more medical images.

SUMMARY OF THE INVENTION
[0011] Before the present system and device and its components are summarized, it is to be understood that this disclosure is not limited to the system and its arrangement as described, as there can be multiple possible embodiments which are not expressly illustrated in the present disclosure. The present disclosure overcomes one or more shortcomings of the prior art and provides additional advantages discussed throughout the present disclosure. Additional features and advantages are realized through the techniques of the present disclosure. It is also to be understood that the terminology used in the description is for the purpose of describing the versions or embodiments only and is not intended to limit the scope of the present application. This summary is not intended to identify essential features of the claimed subject matter nor is it intended for use in detecting or limiting the scope of the claimed subject matter.
[0012] According to embodiments illustrated herein, a method for tissue classification in one or more medical images is disclosed. In one implementation of the present disclosure, the method may involve various steps performed by a processor. The method may involve a step of receiving the one or more medical images. Further, the method may involve a step of pre-processing each of the one or more medical images. Further, the method may involve a step of extracting one or more features from the pre-processed one or more medical images using an encoder of a deep learning model. In an embodiment, the encoder may encode the one or more features into a plurality of blocks. Furthermore, the method may involve a step of semantically segmenting the one or more features from the plurality of blocks to classify each of one or more voxels of the one or more medical images into one or more tissue classifications, using a decoder of the deep learning model. In an embodiment, each tissue classification may be associated with a confidence score. Furthermore, the method may involve a step of post-processing each of the one or more medical images based on the one or more tissue classifications to calculate a final hyperacute mask.
[0013] According to embodiments illustrated herein, a system for tissue classification in the one or more medical images is disclosed. In one implementation of the present disclosure, the system may involve the processor and a memory. The memory is communicatively coupled to the processor. Further, the memory is configured to store one or more executable instructions. Further, the processor may be configured to receive the one or more medical images. Further, the processor may be configured to pre-process each of the one or more medical images. Further, the processor may be configured to extract the one or more features from the pre-processed one or more medical images using the encoder of the deep learning model. In an embodiment, the encoder may encode the one or more features into the plurality of blocks. Furthermore, the processor may be configured to semantically segment the one or more features from the plurality of blocks to classify each of the one or more voxels of the one or more medical images into one or more tissue classifications, using the decoder of the deep learning model. In an embodiment, each tissue classification may be associated with the confidence score. Moreover, the processor may be configured to post-process each of the one or more medical images based on the one or more tissue classifications to calculate the final hyperacute mask.
[0014] According to embodiments illustrated herein, there is provided a non-transitory computer-readable storage medium having stored thereon, a set of computer-executable instructions causing a computer comprising one or more processors to perform various steps. The steps may involve receiving the one or more medical images. Further, the steps may involve pre-processing each of the one or more medical images. Further, the steps may involve extracting the one or more features from the pre-processed one or more medical images using the encoder of the deep learning model. In an embodiment, the encoder may encode the one or more features into the plurality of blocks. Furthermore, the steps may involve semantically segmenting the one or more features from the plurality of blocks to classify each of the one or more voxels of the one or more medical images into one or more tissue classifications, using the decoder of the deep learning model. In an embodiment, each tissue classification may be associated with the confidence score. Moreover, the steps may involve post-processing each of the one or more medical images based on the one or more tissue classifications to calculate the final hyperacute mask.
[0015] The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, examples, and features described above, further aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description.
BRIEF DESCRIPTION OF DRAWINGS
[0016] The detailed description is described with reference to the accompanying figures. In the figures, same numbers are used throughout the drawings to refer like features and components. Embodiments of a present disclosure will now be described, with reference to the following diagrams below wherein:
[0017] Figure 1 illustrates a block diagram describing a system (100) for tissue classification in one or more medical images, in accordance with at least one embodiment of present subject matter.
[0018] Figure 2 illustrates a block diagram showing an overview of various components of an application server (101) configured to classify tissue in the one or more medical images, in accordance with at least one embodiment of present subject matter.
[0019] Figure 3 illustrates a flowchart describing a method (300) for tissue classification in the one or more medical images, in accordance with at least one embodiment of present subject matter.
[0020] Figure 4 illustrates a flowchart (400) describing an exemplary implementation of the method (300) for tissue classification in the one or more medical images, in accordance with at least one embodiment of present subject matter.
[0021] Figure 5 illustrates a block diagram (500) of an exemplary computer system (501) for implementing embodiments consistent with the present subject matter.
[0022] It should be noted that the accompanying figures are intended to present illustrations of exemplary embodiments of the present disclosure. These figures are not intended to limit the scope of the present disclosure. It should also be noted that accompanying figures are not necessarily drawn to scale.
DETAILED DESCRIPTION OF THE INVENTION
[0023] Reference throughout the specification to “various embodiments,” “some embodiments,” “one embodiment,” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, appearances of the phrases “in various embodiments,” “in another embodiments,” “in some embodiments,” “in one embodiment,” or “in an embodiment” in places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the features, structures or characteristics may be combined in any suitable manner in one or more embodiments.
[0024] The words "comprising," "having," "containing," and "including," and other forms thereof, are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items or meant to be limited to only the listed item or items. It must also be noted that, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise. Although any methods similar or equivalent to those described herein can be used in the practice or testing of embodiments of the present disclosure, the exemplary methods are described. The disclosed embodiments are merely exemplary of the disclosure, which may be embodied in various forms.
[0025] The terminology “one or more medical images” and “medical images” has the same meaning and are used alternatively throughout the specification. Further, the terminology “one or more features” and “features” has the same meaning and are used alternatively throughout the specification. Further, the terminology “clinical information” and “clinical data” has the same meaning and are used alternatively throughout the specification. Furthermore, the terminology “plurality of blocks” and “blocks” has the same meaning and are used alternatively throughout the specification. Furthermore, the terminology “one or more voxels” and “voxels” has the same meaning and are used alternatively throughout the specification. Further, the terminology “one or more tissue classifications” and “tissue classifications” has the same meaning and are used alternatively throughout the specification. Further, the terminology “one or more users” and “users” has the same meaning and are used alternatively throughout the specification. Further, the terminology “one or more reference annotations” and “reference annotations” has the same meaning and are used alternatively throughout the specification.
[0026] To address the limitations of conventional imaging-based stroke diagnostic systems, which rely heavily on contrast-enhanced Computed Tomography Perfusion (CTP) and Magnetic Resonance Imaging (MRI) modalities, the present disclosure leverages Non-contrast Computed Tomography (NCCT) imaging scans for ischemic core and penumbra detection. Traditional techniques, while accurate, are inaccessible in many rural and resource-limited settings due to their high costs, need for specialized infrastructure, and use of contrast agents that pose health risks to certain patients. Additionally, existing diagnostic workflows are time-consuming, often delaying critical stroke treatments. The NCCT, despite its broad availability, lacks the sensitivity required for early stroke detection. The disclosed system overcomes these challenges by transforming standard NCCT images into high-utility diagnostic tools through deep learning-based feature extraction and segmentation. This approach eliminates dependency on contrast agents, reduces diagnostic time, and broadens access to advanced stroke care, enabling equitable and efficient clinical intervention even in under-resourced environments.
[0027] Moreover, the present disclosure leverages the ubiquity of the NCCT while enabling accurate estimation of the ischemic core and the penumbra. Thus, an improved system and a structured method capable of analyzing non-contrast images using advanced deep learning techniques would significantly reduce diagnostic latency, mitigate the risks associated with contrast agents, and expand access to high-quality stroke care. Such a solution would empower healthcare providers in underserved regions to deliver timely interventions and improve clinical outcomes for patients suffering from acute ischemic stroke.
[0028] The present disclosure relates to a system and a method for tissue classification in one or more medical images. The system receives one or more medical images, pre-process each of the one or more medical images, and extract one or more features using an encoder of a deep learning model. In an embodiment, the encoder encodes the one or more features into a plurality of blocks. Further, a decoder of the deep learning model semantically segments the one or more features from the plurality of blocks to classify each voxel of the image into one or more tissue classifications, each associated with a confidence score. Furthermore, the system performs post-processing on the classified one or more medical images to calculate a final hyperacute mask.
[0029] Referring to Figure 1 is a block diagram that illustrates the system (100) for tissue classification in the one or more medical images, in accordance with at least one embodiment of the present subject matter. The system (100) typically comprises an application server (101), a database server (102), a communication network (103), and a user computing device (104). The application server (101), the database server (102), and the user computing device (104) are typically communicatively coupled with each other, via the communication network (103). In an embodiment, the application server (101) may communicate with the database server (102), and the user computing device (104) using one or more protocols such as, but not limited to, Hypertext Transfer Protocol (HTTP), Transmission Control Protocol/Internet Protocol (TCP/IP), Wireless Application Protocol (WAP), RF mesh, Bluetooth Low Energy (BLE), and the like, to communicate with one another.
[0030] In an embodiment, the database server (102) may refer to a computing device that may be configured to store, manage, and retrieve data associated with the tissue classification in the one or more medical images. The database server (102) may include a centralized repository specifically configured for storing and maintaining a structured repository of the one or more medical images such as NCCT scans, corresponding to head scans in DICOM (Digital Imaging and Communications in Medicine) or NIfTI (Neuroimaging Informatics Technology Initiative) formats comprising a plurality of axial slices. Additionally, the centralized repository may store the pre-processed medical image data including resized mask slices, normalized intensity values, and combined window stacks with predefined widths and centers. Further, the repository may further maintain extracted multi-scale spatial and contextual features generated by one or more encoders such as MaxViT, Swin Transformer, ViT, CNN-based models, or any transformer or deep learning based architecture. Further, the database server (102) may also store semantically segmented tissue classifications produced by one or more decoders including FPN, U-Net, DeepLab, CNN-based decoders or Transformer-based decoders, and each classification is associated with a confidence score derived from logit masks using a Softmax function. Further, the tissue classifications may include a core infarct tissue, a penumbra tissue, and a normal tissue, each annotated with clinical relevance and stored by the database server (102). Moreover, the centralized repository may contain post-processed hyperacute masks obtained through interpolated mask slices and an Argmax function. Further, the repository may also store training datasets including ground truth annotations from advanced imaging modalities such as CTP and MRI, and historical diagnostic records, thereby enabling continuous improvement and retraining of the deep learning-based tissue classification system.
[0031] The database server (102) enables efficient data retrieval and integration, ensuring seamless processing of the one or more medical images using the encoder and the decoder. In an embodiment, the database server (102) may include hardware and software capable of being realized through various technologies. To implement these capabilities, the database server (102) may utilize a variety of database technologies, including but not limited to relational database management, system (RDBMS), distributed database technology and the like. The database server (102) may also be configured to utilize the application server (101) for storage and retrieval of data required for tissue classification in the one or more medical images.
[0032] A person with ordinary skills in art will understand that the scope of the disclosure is not limited to the database server (102) as a separate entity. In an embodiment, the functionalities of the database server (102) can be integrated into the application server (101) or into the user computing device (104).
[0033] In an embodiment, the application server (101) may refer to a computing device or a software framework hosting an application or a software service. In an embodiment, the application server (101) may be implemented to execute procedures such as, but not limited to, programs, routines, or scripts stored in the database server (102) for supporting the hosted application or the software service. In an embodiment, the hosted application or the software service may be configured to perform one or more predetermined operations. The application server (101) may be realized through various types of application servers such as, but are not limited to, a Java application server, a .NET framework application server, a Base4 application server, a PHP framework application server, or any other application server framework.
[0034] In an embodiment, the application server (101) application may be configured to utilize the database server (102) and the user computing device (104) in conjunction for automated tissue classification and hyperacute stroke assessment based on one or more medical images. In an implementation, the application server (101) corresponds to a computing system that facilitates the coordination and processing of data between the database server (102) and the user computing device (104). The application server (101) manages the flow of data by retrieving the pre-processed one or more medical images, extracted features, tissue classifications, and corresponding confidence scores from the database server (102), and communicates with the user computing device (104) to present the segmented results and hyperacute masks to a user or a healthcare professional. The application server (101) also hosts the necessary software components and deep learning models, including encoders and decoders, to perform end-to-end image segmentation and classification, identifying regions such as the core infarct tissue, the penumbra tissue, and the normal tissue. Further, by integrating multiple deep learning techniques, the application server (101), in conjunction with the database server (102), enables real-time and accurate identification of stroke-affected brain regions, thereby supporting timely and informed deep-learning based clinical decision-making.
[0035] In an embodiment, the application server (101) may be configured to receive the one or more medical images. In an embodiment, the one or more medical images may correspond to non-contrast computed tomography (NCCT) images. Moreover, the NCCT images may correspond to head scan of one or more users.
[0036] In an embodiment, the application server (101) may be configured to pre-process each of the one or more medical images associated with the one or more users.
[0037] In an embodiment, the application server (101) may be configured to extract the one or more features from the pre-processed one or more medical images using the encoder of the deep learning model. In an embodiment, the encoder may encode the one or more features into the plurality of blocks.
[0038] In an embodiment, the application server (101) may be configured to semantically segment the one or more features from the plurality of blocks to classify each of one or more voxels of the one or more medical images into the one or more tissue classifications, using the decoder of the deep learning model. In an embodiment, each of the one or more tissue classification may be associated with the confidence score.
[0039] In an embodiment, the application server (101) may be configured to post-process each of the one or more medical images based on the one or more tissue classifications to calculate the final hyperacute mask.
[0040] In an embodiment, the communication network (103) may correspond to a communication medium through which the application server (101), the database server (102), and the user computing device (104) may communicate with each other. Such communication may be performed in accordance with various wired and wireless communication protocols. Examples of such wired and wireless communication protocols include, but are not limited to, Transmission Control Protocol and Internet Protocol (TCP/IP), User Datagram Protocol (UDP), Hypertext Transfer Protocol (HTTP), Wireless Application Protocol (WAP), File Transfer Protocol (FTP), ZigBee, EDGE, infrared IR), IEEE 802.11, 802.16, 2G, 3G, 4G, 5G, 6G, 7G cellular communication protocols, and/or Bluetooth (BT) communication protocols. The communication network (103) may either be a dedicated network or a shared network. Further, the communication network (103) may include a variety of network devices, including routers, bridges, servers, computing devices, storage devices, and the like. The communication network (103) may include, but is not limited to, the Internet, intranet, a cloud network, a Wireless Fidelity (Wi-Fi) network, a Wireless Local Area Network (WLAN), a Local Area Network (LAN), a cable network, the wireless network, a telephone network (e.g., Analog, Digital, POTS, PSTN, ISDN, xDSL), a telephone line (POTS), a Metropolitan Area Network (MAN), an electronic positioning network, an X.25 network, an optical network (e.g., PON), a satellite network (e.g., VSAT), a packet-switched network, a circuit-switched network, a public network, a private network, and/or other wired or wireless communications network configured to carry data.
[0041] In an embodiment, the user computing device (104) may comprise one or more processors and one or more memory. The one or more memory may store computer-readable instructions that are executable by the one or more processors to interact with the database server (102) and the application server (101) for tissue classification in the one or more medical images. These instructions enable the computing device to receive and display the processed results from the application server (101), including the one or more medical images, clinical information, the final hyperacute mask, and the one or more tissue classifications. The user computing device (104) facilitates the presentation of the processed data, including the clinical information, the one or medical images, the one or more tissue classifications, the final hyperacute mask, to the user or the healthcare professional. The user computing device (104) allows for seamless communication with the application server (101) to receive the calculated final hyperacute mask, enabling users to review and interpret the results. Additionally, the user computing device (104) supports the execution of various user interface applications, making it easier to access, visualize, and understand the insights provided by the system. Through this integration, the user computing device (104) plays an important role in ensuring that healthcare professionals are able to make well-informed decisions based on real-time monitoring of the input one or more medical images, using the encoder and the decoder.
[0042] The system (100) can be implemented using hardware, software, or a combination of both, which includes using where suitable, one or more computer programs, mobile applications, or “apps” by deploying either on-premises over the corresponding computing terminals or virtually over cloud infrastructure. The system (100) may include various micro-services or groups of independent computer programs which can act independently in collaboration with other micro-services. The system (100) may also interact with a third-party or external computer system. Internally, the system (100) may be the central processor of all requests for transactions by the various actors or users of the system.
[0043] Now referring to Figure 2, illustrate a block diagram showing an overview of various components of the application server (101) configured for tissue classification in the one or more medical images, in accordance with at least one embodiment of the present subject matter. Figure 2 is explained in conjunction with elements from Figure 1. In an embodiment, the application server (101) includes a processor (201), a memory (202), a transceiver (203), an input/output unit (204), a user interface unit (205), a pre-processing unit (206), an extracting unit (207), a segmenting unit (208), and a post-processing unit (209). The processor (201) may be communicatively coupled to the memory (202), the transceiver (203), the input/output unit (204), the user interface unit (205), the pre-processing unit (206), the extracting unit (207), the segmenting unit (208), and the post-processing unit (209). The transceiver (203) may be communicatively coupled to the communication network (103) of the system (100).
[0044] The processor (201) comprises suitable logic, circuitry, interfaces, and/or code that may be configured to execute a set of instructions stored in the memory (202), and may be implemented based on several processor technologies known in the art. The processor (201) works in coordination with the memory (202), the transceiver (203), the input/output unit (204), the user interface unit (205), the pre-processing unit (206), the extracting unit (207), the segmenting unit (208), and the post-processing unit (209) for the one or more tissue classifications. Examples of the processor (201) include, but not limited to, a standard microprocessor, microcontroller, central processing unit (CPU), an X86-based processor, a Reduced Instruction Set Computing (RISC) processor, an Application- Specific Integrated Circuit (ASIC) processor, and a Complex Instruction Set Computing (CISC) processor, distributed or cloud processing unit, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions and/or other processing logic that accommodates the requirements of the present invention.
[0045] The memory (202) comprises suitable logic, circuitry, interfaces, and/or code that may be configured to store the set of instructions, which are executed by the processor (201). Preferably, the memory (202) is configured to store one or more programs, routines, or scripts that are executed in coordination with the processor (201). Additionally, the memory (202) may include any computer-readable medium or computer program product known in the art including, for example, volatile memory, such as static random-access memory (SRAM) and dynamic random-access memory (DRAM), and/or non-volatile memory, such as read-only memory (ROM), erasable programmable ROM, a Hard Disk Drive (HDD), flash memories, Secure Digital (SD) card, Solid State Disks (SSD), optical disks, magnetic tapes, memory cards, virtual memory and distributed cloud storage. The memory (202) may be removable, non-removable, or a combination thereof. Further, the memory (202) may include routines, programs, objects, components, data structures, etc., which perform particular tasks or implement particular abstract data types. The memory (202) may include programs or coded instructions that supplement the applications and functions of the system (100). In one embodiment, the memory (202), amongst other things, serves as a repository for storing data processed, received, and generated by one or more of the programs or the coded instructions. In yet another embodiment, the memory (202) may be managed under a federated structure that enables the adaptability and responsiveness of the application server (101).
[0046] The transceiver (203) comprises suitable logic, circuitry, interfaces, and/or code that may be configured to receive, process or transmit information, data or signals, which are stored by the memory (202) and executed by the processor (201). The transceiver (203) is preferably configured to receive, process or transmit, one or more programs, routines, or scripts that are executed in coordination with the processor (201). The transceiver (203) is preferably communicatively coupled to the communication network (103) of the system (100) for communicating all the information, data, signals, programs, routines or scripts through the communication network (103).
[0047] The transceiver (203) may implement one or more known technologies to support wired or wireless communication with the communication network (103). In an embodiment, the transceiver (203) may include but is not limited to, an antenna, a radio frequency (RF) transceiver, one or more amplifiers, a tuner, one or more oscillators, a digital signal processor, a Universal Serial Bus (USB) device, a coder-decoder (CODEC) chipset, a subscriber identity module (SIM) card, and/or a local buffer. Also, the transceiver (203) may communicate via wireless communication with networks, such as the Internet, an Intranet and/or a wireless network, such as a cellular telephone network, a wireless local area network (LAN) and/or a metropolitan area network (MAN). Accordingly, the wireless communication may use any of a plurality of communication standards, protocols and technologies, such as: Global System for Mobile Communications (GSM), Enhanced Data GSM Environment (EDGE), wideband code division multiple access (W-CDMA), code division multiple access (CDMA), time division multiple access (TDMA), Bluetooth, Wireless Fidelity (Wi-Fi) (e.g., IEEE 802.11a, IEEE 802.11b, IEEE 802.11g and/or IEEE 802.11n), voice over Internet Protocol (VoIP), Wi-MAX, a protocol for email, instant messaging, and/or Short Message Service (SMS).
[0048] The input/output (I/O) unit (204) comprises suitable logic, circuitry, interfaces, and/or code that may be configured to receive or present information. The input/output unit (204) comprises various input and output devices that are configured to communicate with the processor (201). Examples of the input devices include but are not limited to, a keyboard, a mouse, a joystick, a touch screen, a microphone, a camera, and/or a docking station. Examples of the output devices include, but are not limited to, a display screen and/or a speaker. The I/O unit (204) may include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like. The I/O unit (204) may allow the system (100) to interact with the user directly or through the user computing devices (104). Further, the I/O unit (204) may enable the system (100) to communicate with other computing devices, such as web servers and external data servers (not shown). The I/O unit (204) can facilitate multiple communications within a wide variety of networks and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite. The I/O unit (204) may include one or more ports for connecting a number of devices to one another or to another server. In one embodiment, the I/O unit (204) allows the application server (101) to be logically coupled to other user computing devices (104), some of which may be built in. Illustrative components include tablets, mobile phones, wireless devices, etc.
[0049] Further, the input/output (I/O) unit (204) may be configured to manage the exchange of data between the application server (101) and the user computing device (104), ensuring that the data is smoothly communicated between the user and the tissue classification system. The I/O unit (204) handles the transmission of the clinical information, the medical images, video information, audio information and the tissue classification results, allowing the user to access real-time reports and contextual insights on the abnormality detection. In an exemplary embodiment, the I/O unit (204) ensures seamless communication by managing data encoding, decoding, and error-checking, and guarantees that the data is delivered in a timely and accurate manner. By enabling reliable and efficient data transfer, the I/O unit (204) optimizes the user experience and supports the system’s overall functionality, enabling healthcare professionals to make quick and informed decisions based on the processed the one or more medical images data.
[0050] Further, the I/O unit (204) may be configured to receive a plurality of information associated with the one or more users. In an embodiment, the plurality of information may include at least one of one or more medical images or a combination thereof. In an exemplary embodiment, the one or more medical images may comprise non-contrast computed tomography (NCCT) images corresponding to head scans of the one or more users. Further, each of the NCCT images may include a plurality of axial slices and may support at least one of the DICOM format or the NIfTI format. In one non-limiting embodiment, the clinical information may be associated with the one or more users received along with the one or more images. In an exemplary embodiment, the clinical information may include at least one of user medical history, radiology reports, electronic health records (EHR), clinical notes, demographic data, diagnostic impressions, or a combination thereof. In another non-limiting embodiment, the one or more users may correspond to patients who has experienced a stroke, a medical condition where the blood supply to the brain is interrupted, causing damage or death of brain cells, or a healthy patient.
[0051] Further, the user interface unit (205) may facilitate interaction between the user and the system (100) by providing a detailed analysis report based on the one or more medical images. The user interface unit (205) enables the one or more users or the healthcare professionals to access the one or more tissue classification outputs along with the final hyperacute mask, which represents a combined delineation of the core infarct and penumbra tissues. In an exemplary embodiment, the interface displays the segmented regions with corresponding confidence scores, allowing visualization of irreversibly damaged tissue and salvageable tissue at risk. In an exemplary embodiment, a report generated may include quantitative assessments and visual overlays on the one or more medical images, thereby aiding the healthcare professionals in making timely decisions regarding patient care and treatment planning. By presenting clear and comprehensive image findings, the user interface unit (205) supports efficient clinical workflows and enhances the accuracy of diagnostic and therapeutic interventions.
[0052] In one embodiment, the application server (101) including the pre-processing unit (206) is disclosed. Further, the pre-processing unit (206) may be configured to pre-process each of the plurality of information associated with the one or more users. Further, the pre-processing unit (206) may be configured for processing the one or more medical images. In an embodiment, processing of the one or more medical images may include resizing each of the one or more medical images into a plurality of mask slices based on one or more predefined dimensions. In an exemplary embodiment, the one or more predefined dimensions may correspond to 32, 256, 256, with an objective to make the images compatible for use in an Artificial Intelligence (AI) model. Furthermore, the pre-processing unit (206) may be configured to split each of the one or more medical images into a plurality of windows. In an embodiment, each window from the plurality of windows may include a distinct predefined window width and center, for example (80,40), (40,40), and (20,20). The pre-processing unit (206) may further combine the plurality of windows by stacking each of the plurality of windows over each other to generate a more detailed input scan. Additionally, the pre-processing unit (206) may normalize one or more intensity values of the one or more medical images to a predefined range. In an exemplary embodiment, the intensity values may be normalized from -1 to 1.
[0053] In one embodiment, the application server (101) including the extracting unit (207) is disclosed. Further, the extracting unit (207) may be configured for extracting the one or more features from the pre-processed one or more medical images using the encoder of the deep learning model. In an exemplary embodiment, the one or more features may correspond to at least one of multi-scale spatial features, contextual features, or a combination thereof. Further, the encoder may be configured to encode the one or more features into the plurality of blocks for subsequent processing. In an exemplary embodiment, the encoder may correspond to at least one of Multi-Axis Vision Transformer (MaxViT), Swin Transformer, Vision Transformer (ViT), Convolutional Neural Network (CNN)-based encoders, or any transformer or deep learning based architecture. Furthermore, the extracting unit (207) may be configured to extract the one or more features from the pre-processed one or more medical images using a combination of convolutional and self-attention operations, thereby enhancing the representational capacity of the encoded features for downstream analysis.
[0054] In another embodiment, the application server (101) including the segmenting unit (208) is disclosed. Further, the segmenting unit (208) may be configured for semantically segmenting the one or more features from the plurality of blocks to classify each of the one or more voxels of the one or more medical images into the one or more tissue classifications using the decoder of the deep learning model. In an embodiment, each of the one or more tissue classifications may be associated with the confidence score. Further, the one or more tissue classifications may include at least one of the core infarct tissue, the penumbra tissue, or the normal tissue. In this context, the core infarct tissue may indicate irreversibly damaged tissue, the penumbra tissue may indicate a salvageable tissue at risk, and the normal tissue may indicate healthy brain regions.
[0055] In an exemplary embodiment, the decoder may correspond to at least one of Feature Pyramid Network (FPN), U-Net decoder, DeepLab decoder, Transformer-based decoders, CNN based decoders, or segmentation decoder. Further, the decoder may be configured to refine the one or more tissue classifications by merging the information from different levels of feature hierarchy and using spatial attention mechanisms. In an exemplary embodiment, the decoder may collate the one or more features extracted by the encoder to classify a particular tissue from the one or more tissue classifications of importance. In an exemplary embodiment, the one or more voxels may refer to the smallest distinguishable three-dimensional units of a medical image, representing volume elements within the scanned anatomical region. Furthermore, associating the confidence score to each of the one or more tissue classifications may include providing a logit mask for each of the one or more voxels corresponding to the one or more tissue classifications, and converting the logit mask into a confidence score by using the Softmax function. In one non-limiting embodiment, the Softmax function may correspond to a mathematical function that converts a set of raw scores or logits into normalized probability values, indicating the relative likelihood of each class in a classification task.
[0056] In one embodiment, the application server (101) including the post-processing unit (209) is disclosed. Further, the post-processing unit (209) may be configured for post-processing each of the one or more medical images based on the one or more tissue classifications to calculate a final hyperacute mask. In an embodiment, the post-processing may include interpolating the plurality of mask slices of the one or more predefined dimensions back to the one or more medical images with an original dimension, and applying the Argmax function to the medical image, post interpolation, to generate a single mask with a plurality of labels. The plurality of labels may correspond to the one or more tissue classifications. In an embodiment, the plurality of labels may include 0 for the normal tissue, 1 for the core infarct tissue, and 2 for the penumbra tissue.
[0057] Further, the post-processing unit (209) may be configured to eliminate false positives in the generated single mask using an island removal technique. In an exemplary embodiment, the Argmax function may be applied along the first dimension of the interpolated logits to consolidate the predictions into a single label per voxel, and the island removal technique is employed to filter out small, isolated regions that do not correspond to true abnormal tissue, thereby refining the accuracy of the final hyperacute mask.
[0058] Furthermore, to provide the final output, the system (100) is configured to generate and present the final hyperacute mask along with the one or more tissue classifications. In an exemplary embodiment, the hyperacute mask may be defined as a combination of the core infarct tissue and the penumbra tissue, where the core infarct tissue represents irreversibly damaged brain regions and the penumbra tissue represents salvageable tissue that is at risk but not yet permanently damaged. The one or more tissue classifications also include the normal tissue, which indicates healthy brain regions.
[0059] In an embodiment, the final output comprising the hyperacute mask and the one or more tissue classifications may be provided to healthcare professionals to assist in clinical decision-making and patient care. Particularly, the identification of the penumbra tissue is critical for initiating timely interventions to salvage the affected regions and prevent further permanent brain damage. This enables healthcare workers to plan and implement appropriate treatment strategies, potentially improving patient outcomes in hyperacute medical conditions such as ischemic stroke.
[0060] In an exemplary embodiment, the system (100) utilizes supervised learning to train the deep learning model. In an exemplary embodiment, one or more reference annotations derived from advanced imaging modalities serve as ground truth. In an exemplary embodiment, the advanced imaging modalities may include at least one of CT Perfusion (CTP) imaging, Magnetic Resonance Imaging (MRI), or a combination thereof. The system (100) may leverage these reference annotations to enable the model to learn to predict corresponding masks on Non-Contrast CT (NCCT) scans.
[0061] In an exemplary embodiment, the supervised learning process may be used to process the NCCT scans using machine learning models trained on annotated datasets of ischemic stroke cases. These models detect subtle changes in tissue density and spatial patterns that correlate with core infarct and penumbra regions, enabling accurate tissue classification even in the absence of advanced imaging. This learning strategy enhances the model’s ability to generalize and detect critical regions using widely available NCCT scans, broadening accessibility to timely stroke diagnosis and intervention. Furthermore, the final hyperacute mask may be generated from the NCCT scan and which may be similar to the MRI or CTP based ground truth mask for all the tissues.
[0062] Now referring to Figure 3, illustrates a flowchart describing the method (300) for tissue classification in the one or more medical images, in accordance with at least one embodiment of the present subject matter. The flowchart is described in conjunction with Figure 1 and Figure 2. The method (300) starts at step (301) and proceeds to step (305).
[0063] In operation, the method (300) may involve a variety of steps, executed by the processor (201), for tissue classification in the one or more medical images.
[0064] At step (301), the method involves receiving the one or more medical images.
[0065] At step (302), the method involves pre-processing each of the one or more medical images.
[0066] At step (303), the method involves extracting the one or more features from the pre-processed one or more medical images using the encoder of the deep learning model. In an embodiment, the encoder may encode the one or more features into the plurality of blocks.
[0067] At step (304), the method involves semantically segmenting the one or more features from the plurality of blocks to classify each of the one or more voxels of the one or more medical images into the one or more tissue classifications, using the decoder of the deep learning model. In an embodiment, each of the one or more tissue classification may be associated with the confidence score.
[0068] At step (305), the method involves post-processing each of the one or more medical images based on the one or more tissue classifications to calculate the final hyperacute mask.
[0069] Let us delve into a detailed working example for the one or more tissue classification in the one or more medical images to calculate the final hyperacute mask according to the present disclosure.
[0070] Working Example 1:
[0071] Consider a stroke patient named ‘ABC’ who arrives at the emergency room exhibiting symptoms of acute ischemic stroke. The hospital’s imaging system receives one or more non-contrast CT (NCCT) scans of the ABC’s brain. These images are pre-processed to correct for noise and standardize intensity levels, preparing them for advanced analysis. The system’s deep learning model then extracts meaningful features from these enhanced images through the encoder, which organizes this information into multiple hierarchical blocks. A decoder within the model performs semantic segmentation on these blocks, classifying each voxel of the medical images into tissue types specifically core infarct (irreversibly damaged tissue), penumbra (salvageable tissue at risk), and normal tissue while assigning confidence scores to each classification.
[0072] Importantly, the classification of each voxel and generation of the final hyperacute mask is time-bound and optimized for real-time application. This ensures that the system can deliver results swiftly, even in high-pressure emergency settings where every minute matters. In the ABC’s case, this rapid classification was crucial. Being from a low-income background, the ABC could not afford an expensive CT Perfusion (CTP) scan, which is traditionally used to distinguish between infarcted and salvageable brain tissue.
[0073] By leveraging only standard NCCT images and the advanced deep learning-based classification system, the ABC’s care team was able to bypass the need for costly CTP imaging. The system generated an accurate hyperacute mask showing a significant region of penumbra, thereby facilitating immediate and confident clinical decision-making. This not only saved the ABC critical treatment time but also substantially reduced the cost of care.
[0074] Based on the hyperacute mask and associated confidence scores, the clinicians initiated timely reperfusion therapy, successfully restoring blood flow and salvaging much of the at-risk brain tissue. As a result, the ABC’s long-term neurological function was preserved, and his rehabilitation period was significantly shortened. The solution provided both clinical value and socioeconomic benefit, demonstrating how the system can improve outcomes while remaining accessible and cost-effective.
[0075] Working Example 2:
[0076] Consider a middle-aged woman named ‘XYZ’ who visits a tertiary hospital with persistent headaches, vision disturbances, and memory lapses. The attending neurologist suspects a high-grade glioma and orders an MRI scan. The hospital’s imaging software, powered by the disclosed method, receives the MRI images specifically a combination of T1-weighted, T2-weighted, and FLAIR sequences.
[0077] These MRI images are pre-processed to remove artifacts, normalize intensities, and align multi-sequence images for unified analysis. The deep learning model then extracts relevant features from the pre-processed MRI scans using its encoder, which structures the extracted features into a plurality of blocks representing various tissue patterns and intensities.
[0078] Using a decoder, the model semantically segments each voxel from these blocks and classifies them into distinct tissue types of tumor core, edema, necrotic region, enhancing tumor, and healthy brain tissue. Each voxel is tagged with a confidence score that reflects the reliability of the classification.
[0079] The classification process is performed in real time and is time-bound, allowing clinicians to obtain a detailed tissue map within minutes of the MRI acquisition. This is crucial for rapidly evolving brain tumors, where early intervention can substantially impact prognosis and treatment options.
[0080] In the XYZ’s case, she was from a rural area with limited access to advanced imaging modalities like MR spectroscopy or PET-MRI, which are often used for detailed tumor assessment but are prohibitively expensive. Thanks to the disclosed method, the team was able to bypass these costly options.
[0081] The post-processing step generated a final hyperacute mask highlighting the exact boundaries and internal composition of the tumor, including regions of active proliferation and necrosis. This allowed the surgical team to plan a precise resection strategy while preserving critical adjacent structures such as the motor cortex and optic pathways.
[0082] The accurate classification also enabled the oncology team to tailor a chemoradiotherapy plan specific to the tumor’s aggressiveness, with real-time confidence scores guiding therapeutic thresholds and margins.
[0083] As a result, the XYZ underwent a successful surgery with minimal post-operative deficits. The early identification of the hyperacute tumor map not only saved her from additional diagnostic costs but also accelerated her treatment schedule, leading to improved clinical outcomes and better quality of life.
[0084] This example illustrates how the disclosed method provides a scalable, cost-effective, and clinically impactful solution for complex conditions like brain cancer, especially in settings with limited resources or financial constraints.
[0085] A person skilled in the art will understand that the scope of the disclosure is not limited to scenarios based on the aforementioned factors and using the aforementioned techniques and that the examples provided do not limit the scope of the disclosure.
[0086] Now referring to Figure 4 illustrates a flowchart (400) describing an exemplary implementation of the method (300) for tissue classification in the one or more medical images, in accordance with at least one embodiment of the present subject matter. The process may begin at step (401) which involves pre-processing of the one or more medical images that consist of 32 mask slices from an NCCT scan. Further, the pre-processing includes windowing the one or more medical images into three different windows using window width and window centre values of (80,40), (40,40), and (20,20), respectively. The one or more medical images of these three windows are subsequently stacked together to obtain a more detailed representation of the one or more medical images.
[0087] At step (402), the encoder is employed to extract multi-scale spatial features and contextual features from the one or more medical images. Further, the encoder utilizes a combination of self-attention and convolutional operations, thereby enabling the encoder to capture both local and global dependencies within the one or more medical images.
[0088] At step (403), the decoder is used to perform semantic segmentation on the extracted features. Further, the decoder receives features from four distinct encoder blocks and generates the logit mask for each of the one or more voxels corresponding to three different tissue classes, the core infarct (irreversibly damaged tissue), the penumbra (salvageable tissue at risk), and the normal tissue across all 32 slices. The decoder refines the segmentation output by merging hierarchical features and employing spatial attention mechanisms.
[0089] At step (404), the Softmax function is applied to the generated logit mask in order to convert the logit masks into confidence probabilities for each of the one or more voxels.
[0090] At step (405), interpolation is performed to restore the spatial resolution of the sliced masks of the one or more medical images from dimensions (32, 256, 256) back to the original shape of the one or more medical images (z, h, w) before the pre-processing of the one or more medical images.
[0091] At step (406), classification of each of the one or more voxels is performed by applying the Argmax function along the class dimension of the confidence probabilities. Further, this results in a single consolidated segmentation mask in which the one or more voxel values are labelled as 0, 1, or 2, corresponding to the normal tissue, the core infarct, and the penumbra, respectively.
[0092] In an exemplary embodiment, the flowchart (400) provides a robust pipeline for accurate and efficient segmentation of stroke regions in the one or more medical images, thereby facilitating improved diagnosis and treatment planning.
[0093] The flowchart (400) outlines the detailed steps involved in acquiring, processing, analysing, and classifying medical image data using advanced encoder-decoder architectures and post-processing techniques to generate clinically meaningful tissue classification outputs.
[0094] Now referring to Figure 5 illustrates a block diagram (500) of an exemplary computer system (501) for implementing embodiments consistent with the present disclosure. Variations of computer system (501) may be used as the method for tissue classification in the one or more medical images. The computer system (501) may comprise a central processing unit (“CPU” or “processor”) (502). The processor (502) may comprise at least one data processor for executing program components for executing the user or the system generated requests. The user may include a person, a person using a device such as those included in this disclosure, or such a device itself. Additionally, the processor (502) may include specialized processing units such as integrated system (bus) controllers, memory management control units, floating point units, graphics processing units, digital signal processing units, or the like. In various implementations the processor (502) may include a microprocessor, such as AMD Athlon, Duron or Opteron, ARM’s application, embedded or secure processors, IBM PowerPC, Intel’s Core, Itanium, Xeon, Celeron or other line of processors, for example. Accordingly, the processor (502) may be implemented using mainframe, distributed processor, multi-core, parallel, grid, or other architectures. Some embodiments may utilize embedded technologies like application-specific integrated circuits (ASICs), digital signal processors (DSPs), or Field Programmable Gate Arrays (FPGAs), for example.
[0095] Processor (502) may be disposed in communication with one or more input/output (I/O) devices via I/O interface (503). Accordingly, the I/O interface (503) may employ communication protocols/methods such as, without limitation, audio, analog, digital, monoaural, RCA, stereo, IEEE-1394, serial bus, universal serial bus (USB), infrared, PS/2, BNC, coaxial, component, composite, digital visual interface (DVI), high-definition multimedia interface (HDMI), RF antennas, S-Video, VGA, IEEE 802.n /b/g/n/x, Bluetooth, cellular (e.g., code-division multiple access (CDMA), high-speed packet access (HSPA+), global system for mobile communications (GSM), long-term evolution (LTE), WiMAX, or the like, for example.
[0096] Using the I/O interface (503), the computer system (501) may communicate with one or more I/O devices. For example, the input device (504) may be an antenna, keyboard, mouse, joystick, (infrared) remote control, camera, card reader, fax machine, dongle, biometric reader, microphone, touch screen, touchpad, trackball, sensor (e.g., accelerometer, light sensor, GPS, gyroscope, proximity sensor, or the like), stylus, scanner, storage device, transceiver, video device/source, or visors, for example. Likewise, an output device (505) may be the user’s smartphone, tablet, cell phone, laptop, printer, fax machine, video display (e.g., cathode ray tube (CRT), liquid crystal display (LCD), light- emitting diode (LED), plasma, or the like), or audio speaker, for example. In some embodiments, a transceiver (506) may be disposed in connection with the processor (502). The transceiver (506) may facilitate various types of wireless transmission or reception. For example, the transceiver (506) may include an antenna operatively connected to a transceiver chip (example devices include the Texas Instruments WiLink WL1283, Broadcom BCM4750IUB8, Infineon Technologies X-Gold 618-PMB9800, or the like), providing IEEE 802.11a/b/g/n, Bluetooth, FM, global positioning system (GPS), and/or 2G/3G/5G/6G HSDPA/HSUPA communications, for example.
[0097] In some embodiments, the processor (502) may be disposed in communication with a communication network (508) via a network interface (507). The network interface (507) is adapted to communicate with the communication network (508). The network interface, coupled to the processor may be configured to facilitate communication between the system and one or more external devices or networks. The network interface (507) may employ connection protocols including, without limitation, direct connect, Ethernet (e.g., twisted pair 10/100/1000 Base T), transmission control protocol/internet protocol (TCP/IP), token ring, or IEEE 802.11a/b/g/n/x, for example. The communication network (508) may include, without limitation, a direct interconnection, local area network (LAN), wide area network (WAN), wireless network (e.g., using Wireless Application Protocol), or the Internet, for example. Using the network interface (507) and the communication network (508), the computer system (501) may communicate with devices such as shown as a laptop (509) or a mobile/cellular phone (510). Other exemplary devices may include, without limitation, personal computer(s), server(s), fax machines, printers, scanners, various mobile devices such as cellular telephones, smartphones (e.g., Apple iPhone, Blackberry, Android-based phones, etc.), tablet computers, eBook readers (Amazon Kindle, Nook, etc.), laptop computers, notebooks, gaming consoles (Microsoft Xbox, Nintendo DS, Sony PlayStation, etc.), or the like. In some embodiments, the computer system (501) may itself embody one or more of these devices.
[0098] In some embodiments, the processor (502) may be disposed in communication with one or more memory devices (e.g., RAM 513, ROM 514, etc.) via a storage interface (512). The storage interface (512) may connect to memory devices including, without limitation, memory drives, removable disc drives, etc., employing connection protocols such as serial advanced technology attachment (SATA), integrated drive electronics (IDE), IEEE-1394, universal serial bus (USB), fiber channel, small computer systems interface (SCSI), etc. The memory drives may further include a drum, magnetic disc drive, magneto-optical drive, optical drive, redundant array of independent discs (RAID), solid-state memory devices, or solid-state drives, for example.
[0099] The memory devices may store a collection of program or database components, including, without limitation, an operating system (516), user interface application (517), web browser (518), mail client/server (519), user/application data (520) (e.g., any data variables or data records discussed in this disclosure) for example. The operating system (516) may facilitate resource management and operation of the computer system (501). Examples of operating systems include, without limitation, Apple Macintosh OS X, UNIX, Unix-like system distributions (e.g., Berkeley Software Distribution (BSD), FreeBSD, NetBSD, OpenBSD, etc.), Linux distributions (e.g., Red Hat, Ubuntu, Kubuntu, etc.), IBM OS/2, Microsoft Windows (XP, Vista/7/8, etc.), Apple iOS, Google Android, Blackberry OS, or the like.
[0100] The user interface (517) is for facilitating the display, execution, interaction, manipulation, or operation of program components through textual or graphical facilities. For example, user interfaces may provide computer interaction interface elements on a display system operatively connected to the computer system (501), such as cursors, icons, check boxes, menus, scrollers, windows, or widgets, for example. Graphical user interfaces (GUIs) may be employed, including, without limitation, Apple Macintosh operating systems’ Aqua, IBM OS/2, Microsoft Windows (e.g., Aero, Metro, etc.), Unix X-Windows, or web interface libraries (e.g., ActiveX, Java, JavaScript, AJAX, HTML, Adobe Flash, etc.), for example.
[0101] In some embodiments, the computer system (501) may implement a web browser (518) stored program component. The web browser (518) may be a hypertext viewing application, such as Microsoft Internet Explorer, Google Chrome, Mozilla Firefox, Apple Safari, or Microsoft Edge, for example. Secure web browsing may be provided using HTTPS (secure hypertext transport protocol), secure sockets layer (SSL), Transport Layer Security (TLS), or the like. Web browsers may utilize facilities such as AJAX, DHTML, Adobe Flash, JavaScript, Java, or application programming interfaces (APIs), for example. In some embodiments the computer system (501) may implement a mail client/server (519) stored program component. The mail server (519) may be an Internet mail server such as Microsoft Exchange, or the like. The mail server may utilize facilities such as ASP, ActiveX, ANSI C++/C#, Microsoft .NET, CGI scripts, Java, JavaScript, PERL, PHP, Python, or WebObjects, for example. The mail server (519) may utilize communication protocols such as internet message access protocol (IMAP), messaging application programming interface (MAPI), Microsoft Exchange, post office protocol (POP), simple mail transfer protocol (SMTP), or the like. In some embodiments, the computer system (501) may implement a mail client (520) stored program component. The mail client (520) may be a mail viewing application, such as Apple Mail, Microsoft Entourage, Microsoft Outlook, or Mozilla Thunderbird.
[0102] In some embodiments, the computer system (501) may store user/application data (521), such as the data, variables, records, or the like as described in this disclosure. Such databases may be implemented as fault-tolerant, relational, scalable, secure databases such as Oracle or Sybase, for example. Alternatively, such databases may be implemented using standardized data structures, such as an array, hash, linked list, struct, structured text file (e.g., XML), table, or as object-oriented databases (e.g., using ObjectStore, Poet, Zope, etc.). Such databases may be consolidated or distributed, sometimes among the various computer systems discussed above in this disclosure. It is to be understood that the structure and operation of the any computer or database component may be combined, consolidated, or distributed in any working combination.
[0103] Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present invention. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer- readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., non-transitory. Examples include Random Access Memory (RAM), Read- Only Memory (ROM), volatile memory, non-volatile memory, hard drives, Compact Disc (CD) ROMs, Digital Video Disc (DVDs), flash drives, disks, and any other known physical storage media.
[0104] In light of the above-mentioned advantages and the technical advancements provided by the disclosed method and system, the claimed steps as discussed above are not routine, conventional, or well understood in the art, as the claimed steps enable the following solutions to the existing problems in conventional technologies. Further, the claimed steps clearly bring an improvement in the functioning of the device itself as the claimed steps provide a technical solution to a technical problem.
[0105] Various embodiments of the disclosure encompass numerous advantages including the system and the method for tissue classification in the one or more medical images. The disclosed method and system have several technical advantages, but are not limited to the following:
• CTP-Equivalent Tissue Classification Using NCCT: The system predicts ischemic core and penumbra masks from non-contrast CT (NCCT) images, eliminating the need for perfusion imaging and enabling comprehensive stroke assessment using a single input modality.
• Accelerated Real-Time Stroke Assessment: The method performs tissue classification with an inference time under five minutes, supporting real-time clinical decision-making for mechanical thrombectomy (MT) eligibility.
• Enhanced Accessibility in Resource-Limited Settings: By exclusively requiring NCCT, the method reduces reliance on costly equipment, contrast agents, and specialized imaging, expanding access to advanced stroke diagnostics in underserved regions.
• Automated Neuroradiologist-Level Segmentation: The semantically segmenting step classifies voxels into tissue classifications with high confidence scores, providing neuroradiologist-level accuracy in identifying ischemic core and penumbra.
• Cost-Efficient Imaging Workflow: The method streamlines stroke assessment by processing a single NCCT series, reducing imaging costs and avoiding the complexities of CTP and MRI-based protocols.
• Optimized Pre-Hospital and In-Hospital Triage: The receiving and pre-processing (302) steps support integration into emergency and stroke center workflows, enabling early, automated identification of ischemic regions and faster patient prioritization.
• Scalable Telemedicine Integration: The system supports deployment at spoke sites and remote diagnostic platforms, enhancing telemedicine-enabled stroke triage and aiding interhospital transfer planning.
• Improved Workflow Efficiency: The post-processing step calculates a final hyperacute mask based on the combination of core tissue and penumbra tissue, reducing imaging-related delays and optimizing time-to-treatment in critical stroke scenarios.
• Quantitative Decision Support: The encoder and decoder architecture provides both visual and quantitative tissue viability insights, aiding clinicians in confident and consistent stroke intervention decisions.
• Robust for Emergency Medical Services: The method is compatible with standard NCCT imaging available in most emergency settings, ensuring scalability and adaptability across diverse clinical environments.

[0106] In summary, the technical advantages of this invention address the challenges associated with conventional stroke imaging workflows, such as the reliance on costly and time-intensive modalities like CT perfusion (CTP) and MRI, which are often unavailable in resource-limited settings. Traditional stroke assessment tools depend on contrast-enhanced imaging and specialized infrastructure, resulting in delayed diagnoses and restricted access to advanced care in underserved regions. In contrast, the disclosed system enables accurate prediction of ischemic core and penumbra directly from the NCCT images using the deep learning-based encoder-decoder architecture. This approach eliminates the need for complex imaging protocols, reduces dependency on contrast agents, and significantly lowers the cost of stroke diagnostics. By semantically segmenting the NCCT images with high confidence scores, the system delivers neuroradiologist-level precision, improving diagnostic consistency and clinician confidence. Furthermore, the system supports real-time decision-making with inference times under five minutes, facilitating rapid triage and eligibility determination for mechanical thrombectomy. Further, the system’s seamless integration into telemedicine platforms and emergency workflows enhances stroke care scalability, optimizes time-to-treatment metrics, and ensures equitable access to life-saving interventions in both urban and rural healthcare settings.
[0107] The claimed invention of the system and the method for tissue classification in the one or more medical images involves tangible components, processes, and functionalities that interact to achieve specific technical outcomes. The system integrates various elements such as processors, memory, databases, classification, confidence scores and informed displaying techniques to effectively perform the tissue classification in the one or more medical images.
[0108] Furthermore, the invention involves a non-trivial combination of technologies and methodologies that provide a technical solution for a technical problem. While individual components like processors, databases, encryption, authorization and authentication are well-known in the field of computer science, their integration into a comprehensive system for tissue classification in the one or more medical images brings about an improvement and technical advancement in the field of clinical data analysis and other related environments.
[0109] In light of the above-mentioned advantages and the technical advancements provided by the disclosed method and system, the claimed steps as discussed above are not routine, conventional, or well understood in the art, as the claimed steps enable the following solutions to the existing problems in conventional technologies. Further, the claimed steps clearly bring an improvement in the functioning of the device itself as the claimed steps provide a technical solution to a technical problem.
[0110] The present disclosure may be realized in hardware, or a combination of hardware and software. The present disclosure may be realized in a centralized fashion, in at least one computer system, or in a distributed fashion, where different elements may be spread across several interconnected computer systems. A computer system or other apparatus adapted for carrying out the methods described herein may be suited. A combination of hardware and software may be a general-purpose computer system with a computer program that, when loaded and executed, may control the computer system such that the computer system carries out the methods described herein. The present disclosure may be realized in hardware that includes a portion of an integrated circuit that also performs other functions.
[0111] A person with ordinary skills in the art will appreciate that the systems, modules, and sub-modules have been illustrated and explained to serve as examples and should not be considered limiting in any manner. It will be further appreciated that the variants of the above disclosed system elements, modules, and other features and functions, or alternatives thereof, may be combined to create other different systems or applications.
[0112] Those skilled in the art will appreciate that any of the aforementioned steps and/or system modules may be suitably replaced, reordered, or removed, and additional steps and/or system modules may be inserted, depending on the needs of a particular application. In addition, the systems of the aforementioned embodiments may be implemented using a wide variety of suitable processes and system modules, and are not limited to any particular computer hardware, software, middleware, firmware, microcode, and the like. The claims can encompass embodiments for hardware and software, or a combination thereof.
[0113] While the present disclosure has been described with reference to certain embodiments, it will be understood by those skilled in the art that various changes may be made, and equivalents may be substituted without departing from the scope of the present disclosure. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the present disclosure without departing from its scope. Therefore, it is intended that the present disclosure is not limited to the particular embodiment disclosed, but that the present disclosure will include all embodiments falling within the scope of the appended claims.
, Claims:WE CLAIM:

1. A method (300) for tissue classification in one or more medical images, wherein the method (300) comprises:
receiving (301), via a processor (201), the one or more medical images;
pre-processing (302), via the processor (201), each of the one or more medical images;
extracting (303), via the processor (201), one or more features from the pre-processed one or more medical images using an encoder of a deep learning model, wherein the encoder encodes the one or more features into a plurality of blocks;
semantically segmenting (304), via the processor (201), the one or more features from the plurality of blocks to classify each of one or more voxels of the one or more medical images into one or more tissue classifications, using a decoder of the deep learning model, wherein each of the one or more tissue classification is associated with a confidence score; and
post-processing (305), via the processor (201), each of the one or more medical images based on the one or more tissue classifications to calculate a final hyperacute mask.

2. The method (300) as claimed in claim 1, wherein the one or more medical images corresponds to non-contrast computed tomography (NCCT) images; wherein the NCCT images corresponds to head scan of one or more users; wherein each image from the one or more medical images comprises a plurality of axial slices; wherein the one or more medical images supports at least one of DICOM (Digital Imaging and Communications in Medicine) format, NIfTI (Neuroimaging Informatics Technology Initiative) format.

3. The method (300) as claimed in claim 1, wherein pre-processing each of the one or more medical images comprises:
resizing each of the one or more medical images into a plurality of mask slices based on one or more predefined dimensions;
splitting each of the one or more medical images into a plurality of windows, wherein each window from the plurality of windows comprises a distinct predefined width and center;
combining the plurality of windows by stacking each of the plurality of windows over each other; and
normalizing one or more intensity values of the one or more medical images to a predefined range.

4. The method (300) as claimed in claim 1, wherein the one or more features corresponds to at least one of multi-scale spatial features, contextual features, or a combination thereof.

5. The method (300) as claimed in claim 1, wherein the encoder corresponds to at least one of Multi-Axis Vision Transformer (MaxViT), Swin Transformer, Vision Transformer (ViT), CNN-based encoders, or any transformer or deep learning based architecture, wherein the encoder extracts the one or more features from the pre-processed one or more medical images using a combination of convolutional and self-attention operations.

6. The method (300) as claimed in claim 1, wherein the decoder corresponds to at least one of Feature Pyramid Network (FPN), U-Net decoder, DeepLab decoder, Transformer-based decoders, CNN based decoders, or segmentation decoder; wherein the decoder refines the one or more tissue classifications by merging information from different levels of feature hierarchy and using spatial attention mechanisms.

7. The method (300) as claimed in claim 1, wherein the one or more tissue classification comprises at least one of a core infarct tissue, a penumbra tissue, or a normal tissue; wherein the core infarct tissue indicates irreversibly damaged tissue, the penumbra tissue indicates salvageable tissue at risk, and the normal tissue indicates healthy brain regions.

8. The method (300) as claimed in claim 1, wherein associating the confidence score to each of the one or more tissue classifications comprises:
providing a logit mask for each of the one or more voxels corresponding to the one or more tissue classifications; and
converting the logit mask into the confidence score by using a Softmax function.

9. The method (300) as claimed in claim 1, wherein the deep learning model is trained using a supervised learning technique; wherein one or more reference annotations derived from advanced imaging modalities serve as ground truth for training the deep learning model; wherein the advanced imaging modalities comprise at least one of CT Perfusion (CTP) imaging, Magnetic resonance imaging (MRI), or a combination thereof.

10. The method (300) as claimed in claim 3, wherein the post-processing comprises:
interpolating the plurality of mask slices of the one or more predefined dimensions back to the one or more medical images with an original dimension; and
applying an Argmax function to a medical image, post interpolation, to get a single mask with a plurality of labels, wherein the plurality of labels corresponds to the one or more tissue classifications.

11. A system (100) for tissue classification in one or more medical images, the system (100) comprises:
a processor (201),
a memory (202) communicatively coupled with the processor (201), wherein the memory (202) is configured to store one or more executable instructions, which cause the processor (201) to:
receive (301), the one or more medical images;
pre-process (302), each of the one or more medical images;
extract (303), one or more features from the pre-processed one or more medical images using an encoder of a deep learning model, wherein the encoder encodes the one or more features into a plurality of blocks;
semantically segment (304), the one or more features from the plurality of blocks to classify each of one or more voxels of the one or more medical images into one or more tissue classifications, using a decoder of the deep learning model, wherein each of the one or more tissue classification is associated with a confidence score; and
post-process (305), each of the one or more medical images based on the one or more tissue classifications to calculate a final hyperacute mask.

12. A non-transitory computer-readable storage medium having stored thereon, a set of computer-executable instructions causing a computer comprising one or more processors to perform steps comprising:
receiving (301), one or more medical images;
pre-processing (302), each of the one or more medical images;
extracting (303), one or more features from the pre-processed one or more medical images using an encoder of a deep learning model, wherein the encoder encodes the one or more features into a plurality of blocks;
semantically segmenting (304), the one or more features from the plurality of blocks to classify each of one or more voxels of the one or more medical images into one or more tissue classifications, using a decoder of the deep learning model, wherein each of the one or more tissue classification is associated with a confidence score; and
post-processing (305), each of the one or more medical images based on the one or more tissue classifications to calculate a final hyperacute mask.

Dated this 28th Day of July 2025

ABHIJEET GIDDE
IN/PA- 4407
AGENT FOR THE APPLICANT