Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network

  Advanced Search  

A roadmap to the integration of early visual modules Abhijit S. Ogale and Yiannis Aloimonos

Summary: A roadmap to the integration of early visual modules
Abhijit S. Ogale and Yiannis Aloimonos
By examining the problem of image correspondence (binocular stereo and optical flow) and its re-
lationship with other modules such as segmentation, shape and depth estimation, occlusion detection,
and local signal processing, we argue that early visual modules are entangled in chicken-and-egg re-
lationships, and unraveling these necessitates a compositional approach. In this paper, we present
compositional algorithms which can match images containing slanted surfaces and images having dif-
ferent contrast, while simultaneously solving other problems as part of the same process. Ultimately,
our goal is to motivate the application of the compositional approach to unify many other early visual
modules. Experimental results have been presented on a large variety of stereo and motion images,
including images with contrast mismatch and images containing untextured slanted surfaces.
1. Introduction
Early vision can be thought of as a collection of modules which deal with the estimation of quantities
such as binocular disparity, optical flow (image motion), texture, occlusions, depth, shape, and various
segmentations. It has long been known that these and other modules are intertwined in a chicken-and-
egg fashion. In this paper, we focus on integrating a small subset of such early modules which are related

A. S. Ogale and Y. Aloimonos are with the Center for Automation Research, University of Maryland, College Park, MD
20742. Email: ogale@cfar.umd.edu, yiannis@cfar.umd.edu


Source: Aloimonos, Yiannis - Center for Automation Research & Department of Computer Science, University of Maryland at College Park


Collections: Engineering