DESC:[0010] Figure 1 illustrates a block diagram of a host device 1000 in communication with an artificial intelligence (AI) health-device 2000 in accordance with the invention. The AI health-device 2000 comprises an AI health-processor 201 in communication with a processor 101 of the host device 1000; characterized in that the AI health-processor 201 is configured to decide, based on a type of medical-image data, a medical condition of the captured medical-image data using a trained algorithm. The AI health-device 2000 also comprises an AI health-memory 202 in communication with the AI health-processor 201. The host device 1000 comprises a memory 102, a medical data capturing means 103 and a user interface 105. The host device 1000 comprises the memory 102 to store the medical data of a patient. The host device 1000 comprises the user interface 105 for the user to interact with the host device 1000.

[0011] In an embodiment, the AI health-device 2000 is removably attached to the host device 1000 via a socket. The AI health-device 2000 comprises the AI health-processor 201 in communication with an AI health-memory 202 embedded in a case. The host device 1000 comprises the socket to receive a plug of the AI health-device 2000. In an embodiment, the AI health-processor 201 is to be plugged by the user within a time window after turning on the host device 1000. In an example, a user of the host device 1000 inserts the plug of AI health-device 2000 into the socket of the host device 1000. The socket is a plug-in slot or an insertion slot of the host device 1000. The plug of the AI health-device 2000 and the socket of the host device 1000 are not explicitly shown in the Figure 1. In an example, the AI health-device 2000 is a device with universal serial bus (USB) based connectivity and the host device 1000 is a high-speed USB 2.0 compliant medical device.

[0012] In another embodiment, the AI health-device 2000 uses protocols to wirelessly connect to the host device 1000. For example, the AI health-device 2000 establishes 802.11 protocol to wirelessly connect to the host device 1000. In an example, the AI health-device 2000 comprises a battery and a user-interface to facilitate the wireless connectivity between the AI health-device and the host device 1000. The battery provides necessary power supply for the operation of the AI health-device 2000.

[0013] In another embodiment, the AI health-processor 201 is implemented as an embedded solution in the host device 1000.

[0014] In an embodiment, the host device 1000 is a medical device used for capturing the medical data of a patient. The medical device is, for example, a fundus camera, a radiography machine, etc. In another embodiment, the host device 1000 is a personal computer, a laptop, a tablet computing device, a personal digital assistant, a smart phone, a mobile phone, a device with an end to end augmented or virtual reality interface, etc. In another embodiment, the host device 1000 is a portable device. In another embodiment, the host device 1000 is a wearable device such as a wrist watch, a smart glasses, etc.

[0015] In an embodiment, the AI health-processor 201 is an integrated circuit (IC). In another embodiment, the AI health-processor 201 is a system-in package (SiP). In another embodiment, the AI health-processor 201 is a system-on-chip (SoC). Once the AI health-processor 201 establishes a communication with the processor 101 of the host device 1000, the processor 101 prompts the user of the host device 1000 to select the medical data associated with the patient. The medical data is, for example, medical text, meta data, medical images or any such other data. The medical images can be one ofa two-dimensional or a three dimensional array of digital image data, a digital video clip, a live stream of a digital medical data, etc. For example, the medical data is a retinal fundus image, a radiographic medical image of a human body part such as a chest X-ray, a mammogram, a dental X-ray, etc., an image or a video of a human body-fluid sample such as a blood sample, a urine sample, a semen sample, etc.

[0016] In an embodiment, the medical data is pre-stored in the memory 102 of the host device 1000. In another embodiment, the host device 1000 comprises the medical data capturing means 103 to capture the medical data associated with the patient. For example, the user of the host device 1000 captures the medical data of the patient using the user interface 105 of the host device 1000. For example, the user captures the medical data of the patient via the user interface 105 via a user input device 1000 such as a mouse, a trackball, a joystick, etc. In another embodiment, the medical data associated with the patient is a live stream of digital information.

[0017] Once the user selects the medical data, the processor 101 transmits the medical data and the type of medical data associated with the medical data to the AI health-processor 201. The AI health-processor 201 receives the medical data and the type of medical data associated with the medical data from the processor 101 of the host device 1000. In an embodiment, the user of the host device 1000 is prompted by the processor 101 to select the type of medical data associated with the selected medical data of the patient via a user interface 105 of the medical device 1000. The type of medical data depicts an organ or a human body part or a human body.

[0018] For example, the host device 1000 is a camera for capturing samples of body fluids like blood, urine, semen, etc. The user is prompted to select the type of body fluid images captured by the host device 1000. This selection is transmitted to the AI health-processor 201 by the processor 101 of the host device 1000.

[0019] The AI health-processor 201 applies the trained algorithm on the medical data of the patient. The AI health-processor 201 comprises the trained algorithm to classify the medical data of the patient. The AI health-processor 201 considers the type of the medical data to classify the medical data of the patient. The type of the medical data indicates one or more body parts present in the medical data and a technology associated with the medical data such as an ultrasound imaging, a radiographic imaging, a tactile imaging, a thermographic imaging, a funduscopic imaging, a positron emission tomography (PET) scan, a magnetic resonance imaging (MRI), etc.

[0020] Trained algorithms of the present invention include algorithms that have been developed using a reference set of known diseased, and normal samples. As used herein, the term “trained algorithm” refers to a class of deep artificial neural network, for example, a convolutional neural network, that can be applied to analyzing visual imagery. The convolutional neural network corresponds to a specific model of an artificial neural network. In an embodiment, one or more convolutional neural networks are applied to process the medical data of the patient. The trained algorithm may also comprise, for example, a support vector machine, recurrent neural networks, deep belief networks, a random forest, gradient boosting, decision trees, boosted decision trees, partial least square classification or regression, branch-and-bound algorithms, neural network models, deep neural networks, convolutional deep neural networks or any combination thereof.

[0021] In an example, the host device 1000 is a fundus camera. The type of the medical data associated with the fundus camera is a retinal-fundus image of the patient. In another example, the host device 1000 is a chest radiographic device. The type of medical data associated with the host device 1000 is a chest radiograph. In another example, the host device 1000 is a dental X-ray machine. The type of the medical data associated with the host device 1000 a dental radiograph.

[0022] The AI health-processor 201 applies the trained algorithm to detect if any abnormalities are present in the medical data. In an embodiment, the AI health-processor 201 also locates anatomical features and/or any artifacts present in the medical data using the trained algorithm. The anatomical features represent one or more body parts of the patient present in the medical data. For example, the anatomical features in a chest radiograph are the lungs, heart, chest wall, great vessels, etc. The artifacts are, for example, metal artifacts. The AI health-processor 201 determines the abnormalities, anatomical features and artifacts based on the type of the medical data. In other words, the AI health-processor 201 gains insight into the body parts focused in the medical data and a technology associated with the medical data from the type of medical data.

[0023] The AI health-processor 201 estimates the medical condition based on the presence of any abnormalities in the medical data using the trained algorithm. The medical condition is one of a healthy condition or a diseased condition. The AI health-processor 201 classifies the medical condition as the diseased condition when the AI health-processor 201 determines the presence of any abnormalities in the medical data. The AI health-processor 201 classifies the medical condition as the healthy condition when the AI health-processor 201 determines an absence of abnormalities in the medical data. The abnormalities in the medical data are based on the type of the medical data.

[0024] The diseased condition represents the presence of one or more diseases. In an embodiment, the AI health-processor 201 may further identify an intensity of the diseased condition using the trained algorithm. The AI health-processor 201 identifies the intensity of each of the one or more diseases present in the medical data. In another embodiment, the AI health-processor 201 determines a size range and a shape of each of the abnormalities present in the medical data. The AI health-processor 201 locates and highlights the size range and the shape of each of the abnormalities in the medical data. This makes further effective analysis of the medical data easier for a medical practitioner.

[0025] In an embodiment, the AI health-processor 201 computes a presence or an absence of one or more medical conditions and a corresponding intensity level of each of the present one or more medical conditions in the medical data based on the type of the medical data using the trained algorithm. In another embodiment, the AI health-processor 201 analyses the medical data to determine a patient-related parameter such as a smoking habit corresponding to the patient using the trained algorithm. For example, the AI health-processor 201 analyses a chest X-ray of a patient to determine the smoking habits of the patient. A smoking addict has specific abnormalities in the chest region. The trained algorithm predicts the smoking habits of the patient based on the detection of the abnormalities developed in the chest region due to smoking.

[0026] The AI health-processor 201 transmits an output of the trained algorithm to the processor 101 of the host device 1000. The output of the trained algorithm is the medical condition of the patient. The output of the trained algorithm may also reflect an overall health condition of the patient. The processor 101 of the host device 1000 displays the output of the trained algorithm via a display of the host device 1000. The processor 101 of the host device 1000 generates a report with the output of the trained algorithm. The report, thus identifies the classification of the medical data as positive or negative for a disease condition associated with the type of the medical data. The report may also comprise the intensity of the disease condition.

[0027] In an example, the report is provided to the user as suitable messages via a pop-up box displayed on a screen of the host device 1000. In another example, the user interface 105 is a smart glasses with augmented reality/virtual reality capabilities. The report is displayed in three-dimensional user interface 105 such as an augmented reality or virtual reality. In an embodiment, the processor 101 communicates the report to the patient via an electronic mail. The processor 101 also stores the report in the memory 102 of the host device 1000 and/or the AI health-memory 202 of the AI health-device 2000.

[0028] In an embodiment, the AI health-processor 201 assesses a quality value of the medical data before locating the medical data. The AI health-processor 201 uses the trained algorithm to assess the quality value of the medical data. The AI health-processor 201 continues with the detection of the abnormalities in the medical data using only when the quality value of the medical data is above a threshold level. The AI health-processor 201 discards the medical data when the quality value of the medical data is below a threshold level. The quality value of the medical data defines an overall grading efficiency of the medical data based on a plurality of quality parameters. The quality parameters are, for example, darkness, light, contrast, color accuracy, tone reproduction, distortion, sharpness, noise, etc. The threshold level is defined by, for example, an annotator during the training of the trained algorithm.

[0029] As used herein, the term “user” is an individual who operates the host device 1000. The terms “user” and “patient” are used interchangeably herein.

[0030] The AI health-processor 201 can thus conveniently measure each of various diseased conditions associated with different body parts of the patient efficiently. Furthermore, the AI health-processor 201 is relatively compact considering the amount of medical data it can analyse. The AI health-processor 201 is portable and easy to handle. This increases flexible reach of medical care to remote places.

[0031] Figure 2 exemplary illustrates a flowchart for analysis of the medical data of the patient in accordance with the invention. At step S1, the user of the host device 1000 removably inserts the AI health-processor 201 into the host device 1000. The host device 1000 comprises the socket to receive the inserted AI health-processor 201. The host device 1000 comprises the processor 101 which is in communication with the AI health-processor 201. In an embodiment, the processor 101 of the host device 1000 transmits the medical data associated with the patient to the AI health-processor 201. Here, the medical data is stored in the memory 102 of the host device 1000. In another embodiment, the host device 1000 comprises the medical data capturing means 103 to capture the medical data associated with the patient. The medical data capturing means 103 is, for example, a camera. The medical data capturing means 103 transmits the captured medical data to the processor 101. The processor 101 further transmits the captured medical data to the AI health-processor 201. The processor 101 also transmits an information regarding the type of the medical data to the AI health-processor 201.

[0032] At step S2, the AI health-processor 201 receives the medical data and the type of the medical data from the processor 101 of the host device 1000. At step S3, the AI health-processor 201 detects the medical condition associated with the medical data of the patient based on the type of the medical data using the trained algorithm. The trained algorithm is, for example, a support vector machine, recurrent neural networks, deep belief networks, a random forest, gradient boosting, decision trees, boosted decision trees, partial least square classification or regression, branch-and-bound algorithms, neural network models, deep neural networks, convolutional deep neural networks or any combination thereof.

[0033] The medical condition is one of the healthy condition or the diseased condition. The AI health-processor 201 classifies the medical condition as the diseased condition when the AI health-processor 201 determines the presence of any abnormalities in the medical data. The AI health-processor 201 classifies the medical condition as the healthy condition when the AI health-processor 201 determines an absence of abnormalities in the medical data. The abnormalities in the medical data are based on the type of the medical data.

[0034] The diseased condition represents the presence of one or more diseases. In an embodiment, the AI health-processor 201 may further identify an intensity of the diseased condition using the trained algorithm. The AI health-processor 201 identifies the intensity of each of the one or more diseases present in the medical data.

[0035] At step S4, the AI health-processor 201 transmits the detected medical condition to the processor 101 of the host device 1000. At step S5, the processor 101 of the host device 1000 displays the detected medical condition associated with patient. The processor 101 generates the report based on the detected medical condition. The processor 101 displays the generated report to the user of the host device 1000 via the user interface 105. In an embodiment, the processor 101 may transmit the generated report to a smart phone using wireless communication. The processor 101 may transmit the generated report to a respective medical practitioner for appropriate medical care of the patient via a network. The network is, for example, a wired network, a wireless network, etc.

[0036] The AI health-processor 201 can be inserted into different host device 1000 to analyse different types of medical data. Since the AI health-processor 201 uses the trained algorithm to analyse different types of medical data, the AI health-processor 201 is compatible for use in different devices, for example, medical device. For example, the AI health-processor 201 can also be used to analyse different types of medical data associated with a patient stored in the memory 102 of the host device 1000. This reduces the cost and time involved in analyzing different types of medical data.

[0037] In an example, the method for analysis of the medical data of a subject is implemented as a software application downloadable by the user on the host device 1000. Here, the host device 1000 is the user’s mobile phone. The AI health-processor 201 is embedded into the mobile phone. The mobile phone comprises the processor 101 which is in communication with the AI health-processor 201 embedded into the mobile phone. The mobile phone comprises the medical data capturing means 103, for example, a camera capable of capturing the medical data associated with the patient. The medical data is a chest radiograph. Once the software is downloaded onto the mobile phone, the mobile phone acts as a stand-alone device capable of analyzing the medical data of the subject. This allows initial medical support at remote areas with no access to a global computer network such as the internet. The medical data capturing means 103 transmits the captured medical data to the processor 101. The processor 101 further transmits the captured medical data to the AI health-processor 201. The processor 101 provides the user with an option to select the type of the medical data via the user interface 105 of the mobile phone. The type of the medical data indicates that the medical data is a radiographic image of the chest of a subject. The processor 101 transmits the type of the medical data to the AI health-processor 201.

[0038] The AI health-processor 201 receives the chest radiograph and the type of the medical data from the processor 101 of the host device 1000. The AI health-processor 201 detects the medical condition associated with the chest radiograph based on the type of the medical data using the trained algorithm. For instance, the trained algorithm is a burned-in code present in a read-only memory of the AI health-processor 201. The AI health-processor 201 transmits the detected medical condition to the processor 101 of the mobile phone. The processor 101 generates the report based on the detected medical condition and displays to the user via the user interface 105.

[0039] The AI health-processor 201 reduces errors resulting from manual identification of various medical conditions during screening of the patient. The AI health-processor 201 acts as an important supporting tool in detecting/monitoring one or more diseases, associated with different body parts of the patient in an effective manner. The AI health-processor 201 reduces the time-consumption involved in a manual recording of the medical condition present/absent in the medical data of the patient.

[0040] The foregoing examples have been provided merely for the purpose of explanation and does not limit the present invention disclosed herein. While the invention has been described with reference to various embodiments, it is understood that the words are used for illustration and are not limiting. Those skilled in the art, may effect numerous modifications thereto and changes may be made without departing from the scope and spirit of the invention in its aspects.
,CLAIMS:We claim:
1. An artificial intelligence (AI) health-device 2000 comprising:
an AI health-processor 201 in communication with a processor 101 of a host device 1000; and
characterized in that
the AI health-processor 201 is configured to decide, based on a type of medical data which may be chosen from the list comprising of Medical text, meta data, medical images; a medical condition of the captured medical data using a trained algorithm.
2. The AI health-device 2000 as claimed in claim 1, wherein the medical condition is one of a healthy condition or a diseased condition.
3. The AI health-device 2000 as claimed in claim 1, wherein the AI health-processor 201 is an integrated circuit, a system in package or the like.
4. The AI health-device 2000 as claimed in claim 1, wherein the type of medical idata depicts an organ or a human body part or a human body.
5. The AI health-device 2000 as claimed in claim 1, wherein the medical images can be be one of two-dimensional digital image data, a three-dimensional digital image data, a digital video image data, a real time image data or the like.
6. A method for analysis of a medical data of a patient, comprising:
? receiving the medical data of the patient and a type of the medical data from a processor 101 of the host device 1000 by an artificial intelligence (AI) health-processor 201 of an AI health-device 2000, wherein the AI health-device 2000 is in communication with the host device 1000;
? detecting a medical condition associated with the medical data based on the type of the medical data by the AI health-processor 201 using a trained algorithm;
? transmitting the detected medical condition to the processor 101 of the host device 1000 by the AI health-processor 201; and
? displaying the detected medical condition by the processor 101 of the host device 1000.
7. The method as claimed in claim 6, wherein the medical condition is one of a healthy condition or a diseased condition.
8. The method as claimed in claim 6, wherein the AI health-processor 201 is an integrated circuit, a system in package or the like.
9. The method as claimed in claim 6, wherein the type of medical data depicts an organ or a human body part or a human body.
10. The method as claimed in claim 6, wherein the medical data may be chosen from the list comprising of Medical text, meta data, medical images. The medical images can be one of a two-dimensional digital image data, a three-dimensional digital image data, a digital video image data, a real time image data or the like.