Title :
A foveated stereo vision system for active depth perception
Author :
Olaya, Emerson J. ; Méndez, Luz A Torres
Author_Institution :
Robot. & Adv. Manuf. Group, Res. Center for Adv. Studies (CINVESTAV) IPN, Ramos Arizpe, Mexico
Abstract :
In this paper we present the design of a biologically inspired active vision system to estimate the depth of objects in a scene. Our design considers geometric characteristics similar to that of human eyes and uses an algorithm that fovealizes the stereo images at points of interest or previous known objects. The proposed algorithm greatly reduces the computational time. We present preliminary simulated results.
Keywords :
active vision; geometry; robot vision; stereo image processing; visual perception; active depth perception; biologically inspired active vision system; foveated stereo vision system; geometric characteristics; object depth estimation; stereo image; Cameras; Eyes; Geometrical optics; Humans; Layout; Machine vision; Robot kinematics; Robot vision systems; Stereo vision; Visual system; active vision system; foveated vision; retinotopic mapping;
Conference_Titel :
Robotic and Sensors Environments, 2009. ROSE 2009. IEEE International Workshop on
Conference_Location :
Lecco
Print_ISBN :
978-1-4244-4777-0
Electronic_ISBN :
978-1-4244-4778-7
DOI :
10.1109/ROSE.2009.5355982
