Title :
Occlusion detection in stereopairs
Author :
Triantafyllidis, G.A. ; Tzovaras, D. ; Strintzis, M.G.
Author_Institution :
Inf. Process. Lab., Aristotelian Univ. of Thessaloniki, Greece
Abstract :
An efficient technique based on the Bayes test is introduced, for detecting occlusions in a noisy stereoscopic image pair. Two hypotheses are used for the formulation of the Bayes decision rules: the first describes the displaced areas while the other describes the occlusion areas. Experimental results illustrating the performance of the proposed techniques are presented and evaluated.
Keywords :
Bayes methods; decision theory; heuristic programming; hidden feature removal; stereo image processing; Bayes decision rules; Bayes test; displaced areas; hypothesis formation; noisy stereoscopic image pair; occlusion detection; performance evaluation; stereopairs; Additive white noise; Cameras; Dynamic programming; Gaussian noise; Geometry; Information processing; Laboratories; Pixel; Stereo vision; Testing;
Conference_Titel :
EUROCON'2001, Trends in Communications, International Conference on.
Conference_Location :
Bratislava, Slovakia
Print_ISBN :
0-7803-6490-2
DOI :
10.1109/EURCON.2001.937773
