Title :
Dense Correspondence with Regional Support for Stereo Vision Systems
Author :
Stivanello, Mauricio Edgar ; Leal, Eduardo Schwinden ; Palluat, Nicolas ; Stemmer, Marcelo Ricardo
Author_Institution :
Dept. of Autom. & Syst. Eng. (DAS), Fed. Univ. of Santa Catarina (UFSC), Florianópolis, Brazil
fDate :
Aug. 30 2010-Sept. 3 2010
Abstract :
In this paper, a new method of stereo dense correspondence based on regional support is proposed. Using the information provided by high frequency components, a new regional point descriptor is introduced to improve matching accuracy. A detailed explanation on how to extend local correspondence algorithms to include the proposed regional support is provided. Results obtained by extending well known algorithms with the proposed approach demonstrate its effectiveness.
Keywords :
image matching; stereo image processing; visual perception; high frequency component; regional point descriptor; stereo dense correspondence; stereo matching; stereo vision system; Calibration; Cameras; Correlation; Cost function; Image reconstruction; Machine vision; Stereo vision; Disparity map; Regional support; Stereo correspondence; Stereo matching; Stereo vision;
Conference_Titel :
Graphics, Patterns and Images (SIBGRAPI), 2010 23rd SIBGRAPI Conference on
Conference_Location :
Gramado
Print_ISBN :
978-1-4244-8420-1
DOI :
10.1109/SIBGRAPI.2010.56
