Method for quantification of uncertainty of contours in manual & auto segmenting algorithms

Abstract

A system ( 10 ) quantifies uncertainty in contours. The system ( 10 ) includes at least one processor ( 42 ) programmed to receive an image ( 18 ) including an object of interest (OOI) ( 20 ). Further, a band of uncertainty ( 32 ) delineating a region ( 34 ) in the received image ( 18 ) is received. The region ( 34 ) includes the boundary of the OOI ( 20 ). The boundary is delineated in the region ( 34 ) using iterative filtering of the region ( 34 ) and a metric of uncertainty of the delineation is determined for the region ( 34 ).

1 . A system for quantification of uncertainty of contours, said system ( 10 ) comprising:
 at least one processor programmed to:
 receive an image including an object of interest (OOI) ( 20 ); 
 receive a band of uncertainty delineating a region in the received image, the region including the boundary of the OOI ( 20 ); 
 delineate the boundary in the region using iterative filtering of the region; and, 
 determine at least one metric of uncertainty of the delineation for the region. 
   
     
     
         2 . The system according to  claim 1 , wherein the band of uncertainty includes an inner contour and an outer contour, the inner contour within the outer contour. 
     
     
         3 . The system according to  claim 1 , wherein the processor is further programmed to:
 display a contour representing the delineated boundary of the OOI, the contour color coded according to metric of uncertainty.   
     
     
         4 . The system according to  claim 1 , wherein the processor is further programmed to:
 for each of at least one of a plurality of sub-regions defining the region:
 iteratively filter the sub-region until the boundary in the sub-region can be delineated with a confidence level exceeding a predetermined level; and, 
 delineate the boundary in the filtered sub-region. 
   
     
     
         5 . The system according to  claim 4 , wherein the processor is further programmed to:
 for each of the at least one of the plurality of sub-regions:
 determine the metric of uncertainty of the delineation for the sub-region, the metric of uncertainty based on the number of iterations. 
   
     
     
         6 . The system according to  claim 4 , wherein the plurality of sub-regions include a sub-region for each point of uncertainty within the region, the sub-region defined by the points of an inner line segment and an outer line segment of the point of uncertainty ( 36 ), the inner line segment spanning from the point of uncertainty to a projection of the point of uncertainty on an inner contour of the band of uncertainty, and the outer line segment spanning from the point of uncertainty to a projection of the point of uncertainty on an outer contour of the band of uncertainty. 
     
     
         7 . The system according to  claim 6 , wherein the projection of the point of uncertainty on the inner contour of the band of uncertainty is a point on the inner contour closest to the point of uncertainty and/or the projection of the point of uncertainty on the outer contour of the band of uncertainty is a point on the outer contour closest to the point of uncertainty. 
     
     
         8 . The system according to  claim 1 , where the processor is further programmed to:
 for each of a plurality of sub-regions defining the region:
 determine whether to filter the sub-region, the determination including at least one of: 
 determining whether the boundary of the OOI can be delineated in the sub-region with a confidence level exceeding a first predetermined level; and, 
 determining whether a stopping condition is met, the stopping condition indicating the confidence level is less than a second predetermined level; 
 in response to determining the the boundary of the OOI can be delineated in the sub-region with the confidence level exceeding the first predetermined level or the stopping condition being met, delineate the boundary of the OOI in the sub-region; and, 
 in response to determining the the boundary of the OOI cannot be delineated in the sub-region with the confidence level exceeding the first predetermined level and the stopping condition not being met, iteratively filter the sub-region a predetermined number of times and repeat the determination as to whether to filter the sub-region. 
   
     
     
         9 . The system according to  claim 8 , wherein the stopping condition is the sub-region having a uniform density. 
     
     
         10 . The system according to  claim 8 , where the processor is further programmed to:
 in response to the stopping condition being met, delineate the boundary of the OOI along the midline of the sub-region.   
     
     
         11 . The system according to  claim 8 , wherein the processor is further programmed to:
 in response to determining the the boundary of the OOI can be delineated in the sub-region with a confidence level exceeding the first predetermined level or the stopping condition being met, determine the metric of uncertainty of the delineation for the sub-region.   
     
     
         12 . A system for quantification of uncertainty of contours, said system ( 10 ) comprising:
 at least one processor programmed to:
 receive an image including an object of interest (OOI); 
 receive a band of uncertainty delineating a region in the received image, the region including the boundary of the OOI; 
 for each of a plurality of sub-regions of the region:
 determine whether to filter the sub-region, the determination including at least one of: 
 determining whether the boundary of the OOI can be delineated in the sub-region with a confidence level exceeding a first predetermined level; and, 
 determining whether a stopping condition is met, the stopping condition indicating the confidence level is less than a second predetermined level; 
 in response to determining the the boundary of the OOI can be delineated in the sub-region with the confidence level exceeding the first predetermined level or the stopping condition being met, delineate the boundary of the OOI in the sub-region; and, 
 in response to determining the the boundary of the OOI cannot be delineated in the sub-region with the confidence level exceeding the first predetermined level and the stopping condition not being met, iteratively filter the sub-region a predetermined number of times and repeat the determination as to whether to filter the sub-region. 
 
   
     
     
         13 . The system according to  claim 12 , wherein the processor is further programmed to:
 in response to determining the the boundary of the OOI can be delineated in the sub-region with a confidence level exceeding the first predetermined level or the stopping condition being met, determine the metric of uncertainty of the delineation for the sub-region.