Local Stereoscopic Depth Estimation Using Ocular Stripe Maps

Abstract

Visual information is represented in the primate visual cortex (area 17, layer 4B) in a peculiar structure of alternating bands of left and right eye dominance. Recently, a number of computational algorithms based on this ocular stripe map architecture have been proposed, from which we selected the cepstral filtering method of Y. Yeshurun & E.L. Schwartz [11] for fast disparity computation due to its simplicity and robustness. The algorithm has been implemented and analyzed. Some special deficiencies have been identified. The robustness against noise and image degradations such as rotation and scaling has been evaluated. We made several improvements to the algorithm. For real image data the cepstral filter behaves like a square autocorrelation of a bandpass filtered version of the original image. The discussed framework is now a reliable single-step method for local depth estimation.