On Modal Modeling for Medical Images Underconstrained Shape Description and Data Compressio.pdf

M.I.T Media Laboratory Perceptual Computing Section Technical Report No. 275 Appeared in Proc. of the IEEE Workshop on Biomedical Image Analysis, Seattle, June 1994 (pp. 70–79). On Modal Modeling for Medical Images: Underconstrained Shape Description and Data Compression Stan Sclaroff and Alex P. Pentland Perceptual Computing Section, The MIT Media Laboratory 20 Ames St., Cambridge MA 02139 Abstract We have previously described modal analysis, an effi- cient, physically-based solution for recovering, tracking, and recognizing solid models from 2-D and 3-D sensor data. The underlying representation consists of two levels: modal deformations, which describe the overall shape of a solid, and displacement maps, which employ a multiscale wavelet representation to provide local and fine surface de- tail. This paper addresses the problem of recovering modal models in the underconstrained case of fitting a 3-D model to contours found in medical slice and X-ray data. We will describe an extension which can be used to incorporate measurement uncertainty while estimating the modal de- formation parameters. Finally, we give details about how to compress dense 3-D point data from surfaces, by use of displacement maps and wavelets. 1 Introduction Contours present an interesting problem in fitting de- formable solids to medical data: when a 3-D model is recovered from contours extracted from 2-D slice data, the model fitting algorithm must make an intelligent guess at what’s going on in the third dimension — things are uncer- tain in the dimension perpendicular to a slice. This classic machine vision problem also arises when fitting partial or occluded data: how can we fill in information where we have no measurements, or noisy measurements? To solve these problems, we should carry measurement uncertainty along in our solid model recovery framework, recovering not only model parameters, but also their distri- butions [7; 27]. Such a framework can be used to set limits on our model