Method for knowledge based image segmentation using shape models

Abstract

A method for segmenting an object of interest from an image of a patient having such object. Each one of a plurality of training shapes is distorted to overlay a reference shape with a parameter Thetai being a measure of the amount of distortion required to effect the overlay. A vector of the parameters Thetai is obtained for every one of the training shapes through the minimization of a cost function along with an estimate of uncertainty for every one of the obtained vectors of parameters Thetai, such uncertainty being quantified as a covariance matrix Sigmai. A statistical model represented as {circumflex over (f)}H (Theta,Sigma) is generated with the sum of kernels having a mean Thetai and covariance Sigmai . The desired object of interest in the image of the patient is identified by positioning of the reference shape on the image and distorting the reference shape to overlay the obtained image with a parameter Theta being a measure of the amount of distortion required to effect the overlay. An uncertainty is quantified as a covariance matrix Sigma and an energy function E=Eshape+Eimage is computed to obtain the probability of the current shape in the statistical shape model Eshape(Theta,Sigma)=-log({circumflex over (f)}H) and the fit in the image Eimage.

Description

[0001] Cross References to Related Applications [0002] This application claims priority to US Provisional Application No. 60 / 698,826, filed July 13, 2005, which application is hereby incorporated by reference. technical field [0003] The present invention relates generally to anatomical object segmentation methods, and more particularly to using prior knowledge of anatomical objects for anatomical object segmentation. Background technique [0004] As is known in the art, many techniques used in quantitative analysis of objects from large three-dimensional (3D) volumes of imaging data, such as CT data, such as anatomy and pathology, involve segmenting objects from neighboring objects . Over the past decade, shape-based segmentation methods have become increasingly common. The Active Shape Model (ASM) and Active Appearance Model (AAM), first introduced in 1995 as described by T.F. Cootes and C.J. Taylor in "Statistical models of appearance for computervision" (Technical R...

AI Technical Summary

Problems solved by technology

However, since modeling is performed after registration, recording errors can be propagated into the model space

Also, the assumption of a Gaussian shape model may be somewhat restrictive

Images