Title :
Adaptive Relaxation Labeling
Author :
Kalayeh, H.M. ; Landgrebe, D.A.
Author_Institution :
Object Recognition Systems, Inc., Princeton, NJ 08540.
fDate :
5/1/1984 12:00:00 AM
Abstract :
Current implementation of probabilistic relaxation labeling (PRL) is based on stationary compatibility coefficients (SCC´s). Such labeling frequently diverges from an achieved minimum labeling error. In this correspondence it is shown that by having nonstationary compatibility coefficients (NSCC´s) the PRL stabilizes about the minimum error which is obtained during the early iterations. Also, a noniterative labeling algorithm which uses NSCC and has a performance similar to that of the modified PRL is suggested.
Keywords :
Classification algorithms; Decision making; Degradation; Image analysis; Information analysis; Labeling; Multispectral imaging; Object recognition; Remote sensing; Smart pixels; Initial labeling probabilities; labeling ambiguity; noniterative adaptive labeling; nonstationary compatibility coefficients; remote sensing;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.1984.4767530
