Title :
Relaxation: Evaluation and Applications
Author :
Fekete, Gyorgy ; Eklundh, Jan-Olof ; Rosenfeld, Azriel
Author_Institution :
Computer Vision Laboratory, Computer Science Center, University of Maryland, College Park, MD 20742.
fDate :
7/1/1981 12:00:00 AM
Abstract :
Probabilistic relaxation labeling processes are iterative parallel schemes that use contextual information to reduce local ambiguities. The behavior of these processes can be described by examining the rates of change and entropies of the probability vectors at each iteration. Examples are given comparing three relaxation processes as applied to several basic image analysis tasks.
Keywords :
Application software; Computer errors; Image analysis; Linear regression; Pattern classification; Pattern recognition; Regression analysis; Statistics; Utility programs; Vectors; Convergence; entropy; performance analysis; relaxation labeling; thresholding;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.1981.4767131
