System and method for ai-based eye condition determinations

Abstract

In some embodiments, a set of eye images related to a subject may be provided to a prediction model. A first prediction may be obtained via the prediction model, where the first prediction is derived from a first eye image and indicates whether an eye condition is present in the subject. A second prediction may be obtained via the prediction model, where the second prediction is derived from a second eye image and indicates that the eye condition is present in the subject. An aspect associated with the first prediction may be adjusted via the prediction model based on the second prediction's indication that the eye condition is present in the subject. One or more predictions related to at least one eye condition for the subject may be obtained from the prediction model, where the prediction model generates the predictions based on the adjustment of the first prediction.

What is claimed is: 
     
         1 . A system for facilitating neural-network-based determinations of eye-related conditions, the system comprising:
 a neural network comprising a plurality of layers, the plurality of layers comprising:
 at least one input layer configured to take, as input, eye images; and 
 at least one hidden layer configured to (i) obtain a first prediction related to a eye condition, (ii) obtain a second prediction related to the eye condition, and (iii) adjust an aspect associated with the first prediction based on an aspect associated with the second prediction, wherein the first prediction is derived from a first eye image of the eye images, and the second prediction is derived from a second eye image of the eye images; and 
   one or more processors executing computer program instructions that, when executed, cause the one or more processors to:
 provide a set of eye images related to a subject to the neural network; and 
 obtain, via the neural network, one or more predictions related to at least one eye condition for the subject, the neural network generating the one or more predictions based on the set of eye images. 
   
     
     
         2 . The system of  claim 1 , further comprising:
 a second neural network configured to perform alignment of eye images,   wherein the one or more processors are caused to:
 provide an initial set of eye images related to the subject to the second neural network to obtain the set of eye images from the second neural network, the set of eye images comprising versions of eye images of the initial set that are aligned by the second neural network, 
 wherein providing the set of eye images to the neural network comprises providing the aligned versions to the neural network such that the neural network generates the one or more predictions based on the aligned versions. 
   
     
     
         3 . The system of  claim 1 , wherein each of the eye images represents a different eye portion of an eye of the subject from one another, and wherein the aspects associated with the first and second predictions respectively comprises a weight or probability associated with the first prediction and a eye location associated with the second prediction such that the at least one hidden layer is configured to perform the adjustment by adjusting the weight or probability associated with the first prediction based on the eye location associated with the second prediction, the eye location associated with the second prediction being a location of the eye that corresponds to the eye portion represented by the second eye image. 
     
     
         4 . The system of  claim 3 , wherein the at least one hidden layer is configured to adjust the weight or probability associated with the first prediction based on a distance between (i) a eye location associated with the first prediction and (ii) the eye location associated with the second prediction, the eye location associated with the first prediction being a location of the eye that corresponds to the eye portion represented by the first eye image. 
     
     
         5 . The system of  claim 4 , wherein the at least one hidden layer is configured to adjust the weight or probability associated with the first prediction based on distances between (i) the eye location associated with the first prediction and (ii) eye locations associated with other predictions derived from at least some of the eye images, the eye locations associated with the other predictions comprising the eye location associated with the second prediction. 
     
     
         6 . The system of  claim 5 , wherein the at least one hidden layer is configured to perform the adjustment by increasing the weight or probability associated with the first prediction based on the eye location associated with the first prediction being close in proximity to the other eye locations associated with the other predictions. 
     
     
         7 . The system of  claim 6 , wherein the at least one hidden layer is configured to perform the adjustment by increasing the weight or probability associated with the first prediction based on (i) the first prediction indicating that the eye condition is present in an eye of the subject, (ii) each of the other predictions indicating that the eye condition is present in the eye of the subject, and (iii) the eye location associated with the first prediction being close in proximity to the other eye locations associated with the other predictions. 
     
     
         8 . The system of  claim 1 , wherein the at least one hidden layer is configured to perform the adjustment by adjusting the aspect associated with the first prediction based on the second prediction indicating that the eye condition is present in an eye of the subject. 
     
     
         9 . The system of  claim 1 , wherein the at least one input layer is configured to take, as input, the eye images and one or more thickness maps, and wherein the neural network is configured to generate the first and second predictions based on the eye images and the one or more thickness maps. 
     
     
         10 . The system of  claim 1 , wherein the one or more processors are caused to:
 provide multiple frames of a same image cut from a eye scan to the neural network to train the neural network, the neural network generating at least one prediction related to at least one eye condition based on the multiple frames of the same image cut.   
     
     
         11 . The system of  claim 10 , wherein providing the multiple frames comprises providing multiple raw frames of the same image cut from the eye scan to the neural network such that each frame of the multiple raw frames of the same image cut comprises different noise patterns from other frames of the multiple raw frames. 
     
     
         12 . The system of  claim 1 , wherein the one or more processors are caused to:
 provide multiple frames of a first image cut from a first eye scan and multiple frames of a second image cut from a second eye scan to the neural network to obtain the one or more predictions,   wherein the neural network generates the one or more predictions based on the multiple frames of the first and second image cuts.   
     
     
         13 . The system of  claim 1 , wherein the one or more predictions relate to at least one of eye ectasia, Keratoconus, corneal graft rejection episode and failure, dry eye syndrome (DES), Fuchs' dystrophy, corneal limbal stem cell deficiency, cataract, or glaucoma. 
     
     
         14 . The system of  claim 1 , wherein the one or more predictions relate to two or more of corneal ectasia, Keratoconus, corneal graft rejection episode and failure, dry eye syndrome (DES), Fuchs' dystrophy, corneal limbal stem cell deficiency, cataract, or glaucoma. 
     
     
         15 . A method implemented by one or more processors executing computer program instructions that, when executed, perform the method, the method comprising:
 providing a set of eye images related to a subject to a prediction model;   obtaining, via the prediction model, a first prediction indicating whether an eye condition is present in the subject, the first prediction being derived from a first eye image of the eye images;   obtaining, via the prediction model, a second prediction indicating that the eye condition is present in the subject, the second prediction being derived from a second eye image of the eye images;   adjusting, via the prediction model, an aspect associated with the first prediction based on the second prediction's indication that the eye condition is present in the subject; and   obtaining, from the prediction model, one or more predictions related to at least one eye condition for the subject, the prediction model generating the one or more predictions based on the adjustment of the first prediction.   
     
     
         16 . The method of  claim 15 , wherein each eye image of the set represents a different eye portion of an eye of the subject from one another, and
 wherein adjusting the aspect associated with the first prediction comprises adjusting a weight or probability associated with the first prediction based on an eye location associated with the second prediction, the eye location associated with the second prediction being a location of the eye that corresponds to the eye portion represented by the second eye image.   
     
     
         17 . The method of  claim 16 , wherein adjusting the weight or probability associated with the first prediction comprises adjusting the weight or probability associated with the first prediction based on a distance between (i) an eye location associated with the first prediction and (ii) the eye location associated with the second prediction, the eye location associated with the first prediction being a location of the eye that corresponds to the eye portion represented by the first eye image. 
     
     
         18 . The method of  claim 15 , wherein adjusting the aspect associated with the first prediction comprises adjusting the aspect associated with the first prediction based on the second prediction indicating that the eye condition is present in an eye of the subject. 
     
     
         19 . The method of  claim 15 , further comprising:
 providing multiple frames of a first image cut from a first eye scan and multiple frames of a second image cut from a second eye scan to the prediction model to obtain the one or more predictions,   wherein the prediction model generates the first prediction based on the multiple frames of the first image cut, and   wherein the prediction model generates the second prediction based on the multiple frames of the second image cut.   
     
     
         20 . Non-transitory computer-readable media comprising computer program instructions that, when executed by one or more processors, perform operations comprising:
 providing a set of eye images related to a subject to a prediction model;   obtaining, via the prediction model, a first prediction indicating whether an eye condition is present in the subject, the first prediction being derived from a first eye image of the eye images;   obtaining, via the prediction model, a second prediction indicating that the eye condition is present in the subject, the second prediction being derived from a second eye image of the eye images;   adjusting, via the prediction model, an aspect associated with the first prediction based on the second prediction's indication that the eye condition is present in the subject; and   obtaining, from the prediction model, one or more predictions related to at least one eye condition for the subject, the prediction model generating the one or more predictions based on the adjustment of the first prediction.