Claims:We claim:
1. An automated system (100) for remotely diagnosing and verifying an eye disease, the method comprising:
enabling, by an image capturing unit (200), to capture eye images of a patient resides in a remote area and storing the captured eye images in a memory (206) associated with the image capturing unit (200);
scanning, by an image scanning unit (202), the memory (206) automatically to determine the presence of the captured eye images;
uploading, by an image transfer unit (204), the determined eye images present in the memory (206) to a cloud server (300);
receiving, by a machine learning model (302), the uploaded eye images present in the cloud server (300) as an input to detect an eye disease associated with the uploaded eye images;
detecting, by the machine learning model (302), glaucoma or retinopathy associated with the uploaded eye images; and
enabling, by the machine learning model (302), to display an analysis of the identified glaucoma or retinopathy associated with the uploaded eye images to an operator present in a rural area and an eye specialist present in an urban area at the same time through a web interface system.
2. The method as claimed in claim 1, further comprising:
enabling, by the machine learning model (302), the eye specialist to review the analysis made by the machine learning model (302); and
enabling, by the machine learning model (302), the eye specialist to approve or reject the analysis made by the machine learning model (302) based on the eye specialist review.
enabling, by the machine learning model (302), the eye specialist to update the reviewed analysis to the cloud server (300).
3. The method as claimed in claim 3, further comprising:
enabling, by the machine learning model (302), to display the reviewed analysis made by the eye specialist in the urban area to the patient resides in the remote area through the web interface.
4. An automated system (100) for remotely diagnosing and verifying an eye disease, the automated system comprising:
an image capturing unit (200) configured to:
capture eye images of a patient resides in a remote area and storing the captured eye images in a memory (206) associated with the image capturing unit (200);
an image scanning unit (202) configured to:
scan the memory (206) automatically to determine the presence of the captured eye images;
an image transfer unit (204) configured to:
upload the determined eye images present in the memory (206) to a cloud server (300);
a machine learning model (302) configured to:
receive the uploaded eye images present in the cloud server as an input to detect an eye disease associated with the uploaded eye images;
detect glaucoma or retinopathy associated with the uploaded eye images; and
display an analysis of the identified glaucoma or retinopathy associated with the uploaded eye images to an operator present in a rural area and an eye specialist present in an urban area at the same time through a web interface system.
5. The automated system (100) as claimed in claim 4, wherein the machine learning model (302) further configure to:
enable the eye specialist to review the analysis made by the machine learning model (302); and
enable the eye specialist to approve or reject the analysis made by the machine learning model based on the eye specialist review.
enable the eye specialist to update the reviewed analysis to the cloud server (300).
6. The automated system (100) as claimed in claim 5, wherein the machine learning model (302) further configure to:
display the reviewed analysis made by the eye specialist in the urban area to the patient resides in the remote area through the web interface system.

, Description:TECHNICAL FIELD
[001] The present invention relates to the eye disease diagnosis and more particularly related to automated system for remote eye disease diagnosis and eye specialist verification using Machine Learning (ML) and Cloud based computing.
BACKGROUND OF INVENTION
[002] The existing Machine Learning (ML) based Eye scan models do not provide the complete eye diagnosis workflow and they fail to enable the doctors to diagnose the patients remotely. Further, the existing ML based Eye scan models requires a technical person to operate a system in remote areas, hence deployment of eye scan ML model in the rural or remote areas is a big challenge. Further, the existing ML based Eye scan models do not incorporate the final human assessment of an eye specialist, who are based in urban area hospitals. Since, the ML based eye diagnosis is not reviewed or signed-off by the human eye specialist and therefore cannot be provided to patient as complete binding eye diagnosis. Thus, the existing Machine Learning model are a failure in terms of actual field deployment.
[003] The above-mentioned shortcomings, disadvantages and problems are addressed herein, and which will be understood by reading and studying the following specification.
OBJECT OF INVENTION
[004] The principal object of the embodiments herein is to provide an automated system and a method for remote eye disease diagnosis and eye specialist verification using Machine Learning (ML) and cloud-based computing.
SUMMARY
[005] Accordingly, the embodiments herein provide an automated system for remotely diagnosing and verifying an eye disease. The method includes capturing eye images of a patient resides in a remote area and storing the captured eye images in a memory associated with an image capturing unit. Further, the method includes scanning the memory automatically to determine the presence of the captured eye images. Further, the method includes uploading the determined eye images present in the memory to a cloud server. Further, the method includes receiving the uploaded eye images present in the cloud server as an input to detect an eye disease associated with the uploaded eye images. Further, the method includes detecting glaucoma or retinopathy associated with the uploaded eye images. Further, the method includes displaying an analysis of the identified glaucoma or Retinopathy associated with the uploaded eye images to an operator present in a rural area and an eye specialist present in an urban area at the same time through a web interface system. Further, the method includes reviewing by the eye specialist present in the urban are, the analysis made by the Machine Learning model and approving or rejecting the analysis made by the Machine Learning model using the Web interface system. Further, the method includes updating by the eye specialist, the reviewed diagnosis to the cloud server and displaying the reviewed diagnosis made by the eye specialist in the urban area to the patient resides in the remote area through the web interface system.
[006] Accordingly, the embodiments herein provide an automated system for remotely diagnosing and verifying an eye disease. The automated system includes an image capturing unit to capture eye images of a patient resides in a remote area and storing the captured eye images in a memory associated with the image capturing unit. Further, the automated system includes an image scanning unit to scan the memory automatically to determine the presence of the captured eye images. Further, the automated system includes an image transfer unit to upload the determined eye images present in the memory to a cloud server. Further, the automated system includes a machine learning (ML) model to receive the uploaded eye images present in the cloud server as an input to detect an eye disease associated with the uploaded eye images. Further, the ML model to detect glaucoma or retinopathy associated with the uploaded eye images. Further, the ML model to display an analysis of the identified glaucoma or Retinopathy associated with the uploaded eye images to an operator present in a rural area and an eye specialist present in an urban area at the same time through a web interface system. Further, the ML model enables the eye specialist present in the urban area to review the analysis made by the ML model. Further, the ML model enables the eye specialist to approve or reject the analysis made by the ML model through the Web interface system and updating the reviewed diagnosis to the cloud server. Further, the ML model displays the reviewed diagnosis made by the eye specialist in the urban area to the patient resides in the remote area through the web interface.
[007] These and other aspects of the embodiments herein will be better appreciated and understood when considered in conjunction with the following description and the accompanying drawings. It should be understood, however, that the following descriptions, while indicating preferred embodiments and numerous specific details thereof, are given by way of illustration and not of limitation. Many changes and modifications may be made within the scope of the embodiments herein without departing from the spirit thereof, and the embodiments herein include all such modifications.
BRIEF DESCRIPTION OF THE DRAWINGS
[008] Embodiments herein are illustrated in the accompanying drawings, throughout which like reference letters indicate corresponding parts in the various figures. The example embodiments herein will be better understood from the following description with reference to the drawings, in which:
[009] FIG. 1 illustrate various units of an automated system for remotely diagnosing and verifying an eye disease, according to an embodiment as disclosed herein;
[0010] FIG. 2 illustrates a web interface displaying patient details, according to an embodiment as disclosed herein;
[0011] FIG. 3 is a flowchart illustrating an automated system and method for remote eye disease diagnosis and eye specialist verification using machine learning (ML) and cloud-based computing, according to an embodiment as disclosed herein;
[0012] FIG. 4 illustrates a TensorFlow Machine Learning model architecture, according to an embodiment as disclosed herein;
[0013] FIG.5 illustrates a Diabetic Retinopathy web interface screen, according to an embodiment as disclosed herein;
[0014] FIG.6 illustrates a Diabetic Retinopathy Artificial intelligence (AI) analysis results web interface screen, according to an embodiment as disclosed herein;
[0015] FIG.7 illustrates a Diabetic Retinopathy eye specialist review screen, according to an embodiment as disclosed herein;
[0016] FIG.8 illustrates a diabetic retinopathy eye specialist review screen, according to an embodiment as disclosed herein;
[0017] FIG.9 illustrates a glaucoma web interface screen, according to an embodiment as disclosed herein;
[0018] FIG.10 illustrates a glaucoma AI analysis results web interface screen, according to an embodiment as disclosed herein; and
[0019] FIG.11 and 12 illustrates a glaucoma eye specialist review screen, according to an embodiment as disclosed herein.
DETAILED DESCRIPTION OF THE INVENTION
[0020] The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. Descriptions of well-known components and processing techniques are omitted, so as not to unnecessarily obscure the embodiments herein. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein can be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
[0021] The embodiments herein provide an automated system and a method to enable an eye specialist based in urban areas to conduct eye camps remotely. The method includes capturing eye images using an image capturing unit (i.e., fundus camera) and feeding the eye images into to the Machine Learning (ML) model based in a cloud server. Further, the method includes conducting analysis of the eye images and provides the analysis of the eyes, whether or not a specific disease such as diabetic retinopathy or glaucoma is present. Further, the method includes displaying the disease information associated with the eyes in a web interface system to review and approve the ML model analysis of the eye images through the web interface system. Further, the method includes combining and displaying an output of the ML model and the review of the eye specialist to a patient resides in remote area through the web interface system. Referring now to the drawings, and more particularly to FIGS. 1 through 12, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments.
[0022] FIG. 1 illustrate various units of an automated system 100 for remotely diagnosing and verifying an eye disease, according to an embodiment as disclosed herein. The automated system 100 includes an image capturing unit 200 and a cloud server 300. Further, image capturing unit 200 includes an image scanning unit 202, an image transfer unit 204 and a memory 206. Further, cloud server 300 includes a machine learning model 302. The image capturing unit 200 configured to capture eye images of a patient resides in a remote area and storing the captured eye images in the memory 206 associated with the image capturing unit 200. The image capturing unit 102 can be at least one of a Fundus eye camera, a mobile camera or any other electronic device having a capability of capturing and storing images. Further, the image scanning unit 202 configured to scan the memory 206 associated with the image capturing unit 200 automatically to determine the presence of the captured eye images. Further, the image transfer unit 204 configured to upload the determined eye images present in the memory 206 to the cloud server 300 through a wireless connection. Further, the machine learning model 302 present in the cloud server 300 configured to receive the uploaded eye images present in the cloud server 300 as an input to detect an eye disease associated with the uploaded eye images. Further, the machine learning model 302 configured to detect glaucoma or retinopathy associated with the uploaded eye images. Further, the machine learning model 302 configured to display an analysis of the identified glaucoma or retinopathy associated with the uploaded eye images to an operator present in a rural area and an eye specialist present in an urban area at the same time through a web interface system (not shown). Further, the machine learning model 302 configured to enable the eye specialist to review the analysis made by the machine learning model and then approve or reject the analysis made by the machine learning model based on the eye specialist review. Further, the machine learning model 302 configured to enable the eye specialist to update the reviewed analysis to the cloud server to display the reviewed analysis made by the eye specialist in the urban area to the patient resides in the remote area through the web interface. Thus, the system 100 helps the Eye specialist located in urban area can review and approve the ML model 302 analysis of the eye images through the web interface system. Finally, the automated system 100 can combine and display an output of the ML model 302 and the review of the eye specialist to patients resides in remote area through the web interface system.
[0023] The memory 206 may include one or more computer-readable storage media. The memory 206 may include non-volatile storage elements. Examples of such non-volatile storage elements may include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories. In addition, the memory 206 may, in some examples, be considered a non-transitory storage medium. The term “non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term “non-transitory” should not be interpreted to mean that the memory 206 is non-movable. In some examples, the memory 206 can be configured to store larger amounts of information than the memory. In certain examples, a non-transitory storage medium may store data that can, over time, change (e.g., in Random Access Memory (RAM) or cache).
[0024] FIG. 1 shows exemplary units of the system 100, but it is to be understood that other embodiments are not limited thereon. In other embodiments, the system 100 may include less or more number of units. Further, the labels or names of the units are used only for illustrative purpose and does not limit the scope of the invention. One or more units can be combined together to perform same or substantially similar function in the system 100.
[0025] FIG. 2 illustrates a web interface system for displaying patient details, according to an embodiment as disclosed herein. The automated system 100 enables the eye specialist based in urban areas to conduct eye camps in rural areas or any other remote setting. The Eye specialist can view the automated eye scans taken by the Fundus eye camera (i.e., Image capturing unit 200) that is located remotely. The Eye images taken by the Fundus camera can be fed into to the Machine Learning (ML) model 302 based in a cloud server 300. Further, the ML model 302 can conduct analysis of the eye images and provides the analysis of the eyes, whether or not a specific disease such as Diabetic Retinopathy or Glaucoma is present. Further, the ML model 302 can display/posts the disease information associated with the eyes in the web interface system. This helps the Eye specialist located in urban area can review and approve the ML model 302 analysis of the eye images through the web interface system. Finally, the automated system 100 can combine and display an output of the ML 302 model and the review of the eye specialist to a patient resides in remote area through the web interface system.
[0026] FIG. 3 is a flowchart 300a illustrating an automated system 100 and method for remote eye disease diagnosis and eye specialist verification using machine learning (ML) and cloud-based computing, according to an embodiment as disclosed herein.
[0027] At step 302a, the method includes capturing eye images of a patient resides in a remote/rural area visiting eye camps and storing the captured eye images in the memory 206 associated with the image capturing unit 200. The memory 206 can be physically integrated with an image capturing unit 200 (for example, a Fundus camera) to store the images taken by the image capturing unit 200. The method allows the image capturing unit to capture the eye images of the patient resides in the remote/rural areas visiting eye camps and storing the captured eye images in the memory 206 associated with the image capturing unit 200.
[0028] At step 304a, the method includes scanning the memory 206 automatically to determine the presence of the captured eye images. The method allows an image scanning unit 202 (i.e., Windows Management Instrumentation (WMI) or WMI events technology) to scan/monitor the memory 206 automatically to determine the presence of the captured eye images.
[0029] At step 306a, the method includes uploading the determined eye images (i.e., new eye images of a patient) present in the memory 206 to the cloud server 300. The method allows the image scanning unit 202 to upload the determined eye images present in the memory 206 to the cloud server 300. The image scanning unit 204 flags the arrival of the new eye images and then a set of eye images are uploaded to cloud sever 300 using an image transfer unit 204.
[0030] At step 308a, the method includes receiving the uploaded eye images present in the cloud server 300 as an input to a machine learning (ML) model 302 to detect an eye disease associated with the uploaded eye images (i.e., an API interface associated with TensorFlow based Machine Learning model deployed 302 in the cloud server 300 is fed with the eye images that are uploaded in the cloud server 300 storage). The method allows the ML model 302 to receive the uploaded eye images present in the cloud server 300 as an input to the ML model 302 to detect the eye disease associated with the uploaded eye images.
[0031] At step 310a, the method includes detecting glaucoma or Retinopathy associated with the uploaded eye images. The method allows the ML model 302 to detect glaucoma or Retinopathy diseases associated with the uploaded eye images. The ML model 302 analyses and predicts the presence of the diseases such as glaucoma or Retinopathy.
[0032] At step 312a, the method includes displaying an analysis of the identified glaucoma or Retinopathy associated with the uploaded eye images to an operator present in the rural area and an eye specialist present in an urban area at the same time through the web interface system. The method allows the machine learning model 302 to enable displaying the analysis of the identified glaucoma or Retinopathy associated with the uploaded eye images to the operator present in the rural area and the eye specialist present in the urban area at the same time through the web interface system (i.e., the eye images and the machine learning model analysis/prediction can be uploaded to the web interface system).
[0033] At step 314a and 316a, the method includes automated system 300 enables the eye specialist present in the urban area to review the analysis made by the ML model 302 and then the eye specialist can approve or rejects the analysis made by the ML model 302 using the Web interface system. Further, the eye specialist can update the reviewed diagnosis to the cloud server 300 and provides/suggests a suitable treatment through the web interface system.
[0034] At step 318a, the method includes displaying the reviewed diagnosis and the suitable treatment made by the eye specialist in the urban area to the patient resides in the remote area through the web interface system using the patient mobile phones or web access points. Thus, the treatment protocol can made available to patients remotely.
[0035] FIG. 4 illustrates a TensorFlow Machine Learning model architecture 400, according to an embodiment as disclosed herein. The TensorFlow Machine Learning model 302 takes the eye image as an input and passes that through several processing layers. These processing layers, namely, Convolutional Layer 1 and Convolutional Layer 2, transform the eye image into output classes. Information from the output classes is classified as Class 0 to Class 9, which provides information on recognizing the original inputs Image.
[0036] The TensorFlow based Machine Learning model 302 analyses the eye images and provides a diagnosis of whether the eye image has Glaucoma or Diabetic Retinopathy or not. The TensorFlow based Machine Learning model 302 compares a new eye image with thousands of eye images that the model has been trained earlier. Thus, the ML model 302 provides the analysis of eye image and sends the scan result through an application program interface (API) integration to the web interface system.
[0037] FIG.5 illustrates a Diabetic Retinopathy web interface screen, according to an embodiment as disclosed herein. In this web interface used by camera operator in the rural area, the eye image of the patient is uploaded via the web interface system. This web interface system uploads the eye image to the cloud 300, where Diabetic Retinopathy (DR) Machine Learning model 302 is deployed for evaluation of eye image. As soon as eye image is uploaded to cloud server 300, Machine Learning model 302 takes that as an input, and starts the image processing and Machine Learning operation on the input image.
[0038] FIG.6 illustrates a Diabetic Retinopathy Artificial intelligence (AI) analysis results web interface screen, according to an embodiment as disclosed herein. Once the Image is uploaded to the cloud 300 in the previous stage, the Machine Learning model 302 processes the image and forwards the Image classification information to the web system. This screen on the web system displays the output of Machine learning model 302. This screen is accessible by the operator in field and rural area, as well as the doctor or specialist in urban area, and hence the system enables doctor to review the AI analysis and provide their respective inputs on the web system.
[0039] FIG.7 illustrates a Diabetic Retinopathy eye specialist review screen, according to an embodiment as disclosed herein. In an embodiment, the eye specialist can review all the patients, which so far have gone through the image scan and analysis. The eye specialist can download and view the eye images, analyze the image, review the output by Diabetic Retinopathy Machine Learning model 300 and input their remarks to the system. The specialist remarks can then be viewed by operator in rural or remote area, and hence the operator can provide the remarks to the patient who is going through the entire eye scan process within a short period of time.
[0040] FIG.8 illustrates a diabetic retinopathy eye specialist review screen, according to an embodiment as disclosed herein. The diabetic retinopathy eye specialist review screen shows the eye image that are taken in a remote area and then can be reviewed and processed by the eye specialist in the urban location.
[0041] FIG.9 illustrates a glaucoma web interface screen, according to an embodiment as disclosed herein. The glaucoma web interface screen shown can be used by the camera operator in the rural area or remote location. The eye image of the patient is uploaded via the web interface system. This web interface system uploads the eye image to the cloud 300, where Glaucoma Machine Learning model 302 is deployed for evaluation of eye image. As soon as eye image is uploaded to cloud server 303, the machine learning model 302 takes that as an input and starts the image processing operation on the input image.
[0042] FIG.10 illustrates a glaucoma AI analysis results web interface screen, according to an embodiment as disclosed herein. Once the captured eye Images are uploaded to the cloud server 300 in the previous stage, the Glaucoma Machine Learning model 302 processes the image and forwards the Image classification information to the web interface system. This screen on the web interface system displays the output of Machine Learning model 302. This screen is accessible by the operator in field and rural area, as well as the doctor or specialist in urban area, and hence the system 100 enables doctor to review the AI analysis and provide their respective inputs on the web system.
[0043] FIG.11 and 12 illustrates a glaucoma eye specialist review screen, according to an embodiment as disclosed herein. The glaucoma eye specialist review screen helps the eye specialist to review all the patients, which so far have gone through the image scan and analysis. The eye specialist can download and view the eye images, analyze the image, review the output by Glaucoma Machine Learning model 302 and input their remarks to the system 100. The specialist remarks can then be viewed by operator in rural or remote area, and hence the operator can provide the remarks to the patient who is going through the entire eye scan process within a short period of time. FIG. 12 shows the eye image that is taken in a remote area and then can be reviewed and processed by the eye specialist in urban location.
[0044] The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the spirit and scope of the embodiments as described herein.