Title :
Estimation of discontinuous displacement vector fields with the minimum description length criterion
Author_Institution :
Dept. of Biophys., German Cancer Res. Center, Heidelberg, Germany
Abstract :
A noniterative approach to determine displacement vector fields with discontinuities is described. In order to overcome the limitations of current methods, the problem is regarded as a general modeling problem. From this point of view the imaging field consists of a set of regions with common properties and their boundaries. The strategy proposed is an analysis of consistency of the displacement estimators between different levels of regularization. This gives local but noisy evidence of possible model boundaries at multiple scales. With the two constraints of continuous lines of discontinuities and the spatial coincidence assumption, consistent boundary evidence is found. Based on this combined evidence the model is updated so that it describes homogeneous regions with sharp discontinuities.<>
Keywords :
computer vision; computerised picture processing; consistency; discontinuities; displacement vector fields; general modeling problem; homogeneous regions; imaging field; minimum description length criterion; noniterative approach; regularization; spatial coincidence assumption; Biophysics; Cancer; Computer vision; Displacement measurement; Geometry; Image segmentation; Least squares approximation; Motion analysis; Stereo vision; Uncertainty;
Conference_Titel :
Computer Vision and Pattern Recognition, 1991. Proceedings CVPR '91., IEEE Computer Society Conference on
Conference_Location :
Maui, HI, USA
Print_ISBN :
0-8186-2148-6
DOI :
10.1109/CVPR.1991.139702