Systems and methods for automated detection of regions of interest in retinal images

Abstract

Embodiments disclose systems and methods that aid in screening, diagnosis and/or monitoring of medical conditions. The systems and methods may allow, for example, for automated identification and localization of lesions and other anatomical structures from medical data obtained from medical imaging devices, computation of image-based biomarkers including quantification of dynamics of lesions, and/or integration with telemedicine services, programs, or software.

What is claimed is: 
     
         1 . A computing system for automated detection of regions of interest in retinal images, the computing system comprising:
 one or more hardware computer processors; and   one or more storage devices configured to store software instructions configured for execution by the one or more hardware computer processors in order to cause the computing system to:
 access a retinal image; 
 extract regions of interest with one or more requested desired properties from the retinal image using multi-scale morphological filterbank analysis; 
 store a binary image representing the regions of interest in a memory of the computing system. 
   
     
     
         2 . The computing system of  claim 1 , wherein the requested desired property is area of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have areas within one or more desired ranges of areas. 
     
     
         3 . The computing system of  claim 1 , wherein the requested desired property is color of a local region, defined as one or more of mean, median, or an order statistic value of colors of pixels in the local region, and the request comprises retaining only those local regions from a plurality of local regions which have colors within one or more desired ranges of colors. 
     
     
         4 . The computing system of  claim 1 , wherein the requested desired property is eccentricity of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have eccentricities within one or more desired ranges of eccentricities. 
     
     
         5 . The computing system of  claim 1 , wherein the requested desired property is blobiness of a local region, defined as one or more of mean, minimum, maximum, median, or any order statistic value of one or more of: determinant of a Hessian matrix, or ratio of Eigen values of a Hessian matrix corresponding to each pixel within the local region, and the request comprises retaining only those local regions from a plurality of local regions which have blobiness within one or more desired ranges of blobiness. 
     
     
         6 . The computing system of  claim 1 , wherein the requested desired property is contrast of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have contrasts within one or more desired ranges of contrasts. 
     
     
         7 . The computing system of  claim 1 , wherein the requested desired property is vesselness of a local region, defined as one or more of mean, median, or any order statistic value of vesselness of pixels in the local region, and the request comprises retaining only those local regions from a plurality of local regions which have vesselness within one or more desired ranges of vesselness. 
     
     
         8 . The computing system of  claim 1 , wherein the requested desired properties are one or more of area, color, contrast, eccentricity, blobiness, or vesselness of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have the above desired properties within one or more desired ranges wherein the ranges are defined separately for each property. 
     
     
         9 . The computing system of  claim 1 , wherein the morphological filterbank is a median filterbank. 
     
     
         10 . The computing system of  claim 1 , wherein the computing system is further configured to resize the retinal image to a standard size wherein the standard size could be determined using:
 a fixed value;   field of view information provided or inferred from a structure in the retinal image;   a diameter of a fundus mask.   
     
     
         11 . The computing system of  claim 1 , wherein the computing system is further configured to:
 scale the retinal image progressively up or down using a set of scaling factors;   designate local regions within a retinal image as regions of interest; and   include the regions of interest from each scale as regions of interest across multiple scales.   
     
     
         12 . The computing system of  claim 1 , wherein the computing system is further configured to automatically perform at least one of lesion localization, screening, image quality assessment, and analysis of a vascular network. 
     
     
         13 . A computer-implemented method for automated detection of regions of interest in retinal images, the method comprising:
 as implemented by one or more computing devices configured with specific executable instructions:
 accessing a retinal image; 
 extracting regions of interest with one or more requested desired properties from the retinal image using multi-scale morphological filterbank analysis; 
 storing a binary image representing the regions of interest in a memory. 
   
     
     
         14 . The computer-implemented method of  claim 13 , wherein the requested desired property is area of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have areas within one or more desired ranges of areas. 
     
     
         15 . The computer-implemented method of  claim 13 , wherein the requested desired property is color of a local region, defined as one or more of mean, median, or an order statistic value of colors of pixels in the local region, and the request comprises retaining only those local regions from a plurality of local regions which have colors within one or more desired ranges of colors. 
     
     
         16 . The computer-implemented method of  claim 13 , wherein the requested desired property is eccentricity of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have eccentricities within one or more desired ranges of eccentricities. 
     
     
         17 . The computer-implemented method of  claim 13 , wherein the requested desired property is blobiness of a local region, defined as one or more of mean, minimum, maximum, median, or any order statistic value of one or more of: determinant of a Hessian matrix, or ratio of Eigen values of a Hessian matrix corresponding to each pixel within the local region, and the request comprises retaining only those local regions from a plurality of local regions which have blobiness within one or more desired ranges of blobiness. 
     
     
         18 . The computer-implemented method of  claim 13 , wherein the requested desired property is contrast of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have contrasts within one or more desired ranges of contrasts. 
     
     
         19 . The computer-implemented method of  claim 13 , wherein the requested desired property is vesselness of a local region, defined as one or more of mean, median, or any order statistic value of vesselness of pixels in the local region, and the request comprises retaining only those local regions from a plurality of local regions which have vesselness within one or more desired ranges of vesselness. 
     
     
         20 . The computer-implemented method of  claim 13 , wherein the requested desired properties are one or more of area, color, contrast, eccentricity, blobiness, or vesselness of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have the above desired properties within one or more desired ranges wherein the ranges are defined separately for each property. 
     
     
         21 . The computer-implemented method of  claim 13 , wherein the morphological filterbank is a median filterbank. 
     
     
         22 . The computer-implemented method of  claim 13 , further comprising resizing the retinal image to a standard size wherein the standard size could be determined using:
 a fixed value;   field of view information provided or inferred from a structure in the retinal image;   a diameter of a fundus mask.   
     
     
         23 . The computer-implemented method of  claim 13 , further comprising:
 scaling the retinal image progressively up or down using a set of scaling factors;   designating local regions within a retinal image as regions of interest; and   including the regions of interest from each scale as regions of interest across multiple scales.   
     
     
         24 . The computer-implemented method of  claim 13 , further comprising automatically performing at least one of lesion localization, screening, image quality assessment, and analysis of a vascular network. 
     
     
         25 . Non-transitory computer storage that stores executable program instructions that, when executed by one or more computing devices, configure the one or more computing devices to perform operations comprising:
 accessing a retinal image;   extracting regions of interest with one or more requested desired properties from the retinal image using multi-scale morphological filterbank analysis;   storing a binary image representing the regions of interest in a memory.   
     
     
         26 . The non-transitory computer storage of  claim 25 , wherein the requested desired property is area of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have areas within one or more desired ranges of areas. 
     
     
         27 . The non-transitory computer storage of  claim 25 , wherein the requested desired property is color of a local region, defined as one or more of mean, median, or an order statistic value of colors of pixels in the local region, and the request comprises retaining only those local regions from a plurality of local regions which have colors within one or more desired ranges of colors. 
     
     
         28 . The non-transitory computer storage of  claim 25 , wherein the requested desired property is eccentricity of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have eccentricities within one or more desired ranges of eccentricities. 
     
     
         29 . The non-transitory computer storage of  claim 25 , wherein the requested desired property is blobiness of a local region, defined as one or more of mean, minimum, maximum, median, or any order statistic value of one or more of: determinant of a Hessian matrix, or ratio of Eigen values of a Hessian matrix corresponding to each pixel within the local region, and the request comprises retaining only those local regions from a plurality of local regions which have blobiness within one or more desired ranges of blobiness. 
     
     
         30 . The non-transitory computer storage of  claim 25 , wherein the requested desired property is contrast of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have contrasts within one or more desired ranges of contrasts. 
     
     
         31 . The non-transitory computer storage of  claim 25 , wherein the requested desired property is vesselness of a local region, defined as one or more of mean, median, or any order statistic value of vesselness of pixels in the local region, and the request comprises retaining only those local regions from a plurality of local regions which have vesselness within one or more desired ranges of vesselness. 
     
     
         32 . The non-transitory computer storage of  claim 25 , wherein the requested desired properties are one or more of area, color, contrast, eccentricity, blobiness, or vesselness of a local region, and the request comprises retaining only those local regions from a plurality of local regions which have the above desired properties within one or more desired ranges wherein the ranges are defined separately for each property. 
     
     
         33 . The non-transitory computer storage of  claim 25 , wherein the morphological filterbank is a median filterbank. 
     
     
         34 . The non-transitory computer storage of  claim 25 , further comprising resizing the retinal image to a standard size wherein the standard size could be determined using:
 a fixed value;   field of view information provided or inferred from a structure in the retinal image;   a diameter of a fundus mask.   
     
     
         35 . The non-transitory computer storage of  claim 25 , further comprising:
 scaling the retinal image progressively up or down using a set of scaling factors;   designating local regions within a retinal image as regions of interest; and   including the regions of interest from each scale as regions of interest across multiple scales.   
     
     
         36 . The non-transitory computer storage of  claim 25 , further comprising: automatically performing at least one of lesion localization, screening, image quality assessment, and analysis of a vascular network.