System and method for ultrasound analysis

Abstract

An exemplary system, method and computer-accessible medium for detecting an anomaly(ies) in an anatomical structure(s) of a patient(s) includes receiving imaging information related to the anatomical structure(s) of the patient(s), classifying a feature(s) of the anatomical structure(s) based on the imaging information using a neural network (s), and detecting the anomaly(ies) based on data generated using the classification procedure.

What is claimed is: 
     
         1 . A system for analyzing at least one anatomical structure of at least one patient, comprising:
 at least one neural network trained on multiple recognition and/or analysis procedures, trained in order of their difficulty level; and   a specifically configured computer hardware arrangement configured to:   receive imaging information related to the at least one anatomical structure of the at least one patient;   using said at least one neural network, produce at least one non-procedure specific feature of the at least one anatomical structure based on the imaging information; and   perform one of said recognition and/or analysis procedures using said non-procedure-specific features to determine at least one anomaly or as input to a further procedure.   
     
     
         2 . The system of  claim 1 , wherein the imaging information includes at least three images of the at least one anatomical structure. 
     
     
         3 . The system of  claim 1 , wherein the imaging information includes ultrasound imaging information. 
     
     
         4 . The system of  claim 1 , wherein the at least one anatomical structure is a heart and wherein said multiple recognition and/or analysis procedures comprise:
 a view detection procedure to detect a view of a particular imaging frame in the imaging information;   a systole/diastole detection procedure;   a part segmentation procedure to segment parts of the heart of the at least one patient from a background;   a valve localization procedure to localize a heart valve; and   an anomaly detection procedure.   
     
     
         5 . The system of  claim 1 , wherein the at least one state of said heart includes at least one of (i) a systole state of a heart of the at least one patient, (ii) a diastole state of the heart of the at least one patient, (iii) an inflation state of the heart of the at least one patient or (iv) a deflation state of the heart of the at least one patient. 
     
     
         6 . The system of  claim 1 , wherein the further procedure is configured to determine an ejection fraction using output of said view detection and part segmentation procedures, where the segmentation segments the left ventricle. 
     
     
         7 . The system of  claim 6 , wherein the further procedure is configured to place at least one Doppler measuring point within an appropriate view of said heart. 
     
     
         8 . The system of  claim 1 , wherein the imaging information includes a plurality of images at different resolutions, and wherein the at least one neural network includes a plurality of neural networks, each of the neural networks being associated with one of the images. 
     
     
         9 . A method for analyzing at least one anatomical structure of at least one patient, comprising:
 training at least one neural network on multiple recognition and/or analysis procedures, trained in order of their difficulty level;   receiving imaging information related to the at least one anatomical structure of the at least one patient;   using said at least one neural network, producing at least one non-procedure specific feature of the at least one anatomical structure based on the imaging information; and   performing one of said recognition and/or analysis procedures using said non-procedure-specific features to determine at least one anomaly or as input to a further procedure.   
     
     
         10 . The method of  claim 9 , wherein the imaging information includes at least three images of the at least one anatomical structure. 
     
     
         11 . The method of  claim 9 , wherein the imaging information includes ultrasound imaging information. 
     
     
         12 . The method of  claim 9 , wherein the at least one anatomical structure is a heart and wherein said multiple recognition and/or analysis procedures comprise:
 a view detection procedure to detect a view of a particular imaging frame in the imaging information;   a systole/diastole detection procedure;   a part segmentation procedure to segment parts of the heart of the at least one patient from a background;   a valve localization procedure to localize a heart valve; and   an anomaly detection procedure.   
     
     
         13 . The method of  claim 9 , wherein the at least one state of said heart includes at least one of (i) a systole state of a heart of the at least one patient, (ii) a diastole state of the heart of the at least one patient, (iii) an inflation state of the heart of the at least one patient or (iv) a deflation state of the heart of the at least one patient. 
     
     
         14 . The method of  claim 9 , wherein the further procedure comprises determining an ejection fraction using output of said view detection and part segmentation procedures, where the segmentation segments the left ventricle. 
     
     
         15 . The method of  claim 12 , wherein the further procedure comprises placing at least one Doppler measuring point within an appropriate view of said heart. 
     
     
         16 . The method of  claim 9 , wherein the imaging information includes a plurality of images at different resolutions, and wherein the at least one neural network includes a plurality of neural networks, each of the neural networks being associated with one of the images. 
     
     
         17 . The system of  claim 6 , wherein the further procedure is configured to determine three dimensional volumes of the systole and diastole of the left ventricle and to fit a three dimensional mesh to each of said volumes. 
     
     
         18 . The system of  claim 6 , wherein the further procedure is configured to determine said ejection fraction from said meshes. 
     
     
         19 . The system of  claim 4 , wherein the further procedure is configured to generate measurements of said heart, including at least one of: a dimension of the left ventricle in systole and diastole, right ventricular assessment, LA size, measurement of the aortic valve annulus, the aortic sinuses, the ascending aorta, the pulmonary valve, the mitral valve annulus and the tricuspid valve annulus. 
     
     
         20 . The method of  claim 14 , wherein the further procedure comprises determining three dimensional volumes of the systole and diastole of the left ventricle and to fit a three dimensional mesh to each of said volumes. 
     
     
         21 . The method of  claim 20 , wherein the further procedure comprises determining said ejection fraction from said meshes. 
     
     
         22 . The method of  claim 12 , wherein the further procedure comprises generating measurements of said heart, including at least one of: a dimension of the left ventricle in systole and diastole, right ventricular assessment, LA size, measurement of the aortic valve annulus, the aortic sinuses, the ascending aorta, the pulmonary valve, the mitral valve annulus and the tricuspid valve annulus.