Abstract
A model is described for stereopsis. Two stereo images are bandpass filtered at multiple scales using a Laplacian pyramid data structure.1 These bandpass image pairs are then matched in parallel at all scales and locations using an iterative, relaxation labelinglike process. At a given spatial location i and scale (or pyramid level) k, the state consists of a disparity estimate (the disparity mean Mik) and a confidence in that estimate (the disparity variance Vik). Disparity matches are only made near the disparity currently indicated, Mik. The update of Mik includes a data-driven term (from the matches), a local averaging term (from Mjk, where j is a location near i in the image at scale k), a top-down term (from Mik+1), and a bottom-up term (from Mik−1). These terms are gated by the corresponding confidence terms (Vik, Vik+1, and Vik−1), which reflect the quality of the match. Simulations of the model are discussed both in terms of the model’s efficacy in computing stereoscopic depth, and in general concerning how computation may take place cooperatively across multiple scales.
© 1986 Optical Society of America
PDF ArticleMore Like This
Clifton M. Schor
MB3 OSA Annual Meeting (FIO) 1986
Antonio Medina Puerta
TuY18 OSA Annual Meeting (FIO) 1990
J. Siderov and R. S. Harwerth
FKK5 OSA Annual Meeting (FIO) 1990