Layer extraction from multiple images containing reflections and transparency

Abstract
Many natural images contain reflections and transparency, i.e., they contain mixtures of reflected and transmitted light. When viewed from a moving camera, these appear as the su- perposition of component layer images moving relative to each other. The problem of multiple motion recovery has been pre- viously studied by a number of researchers. However, no one has yet demonstrated how to accurately recover the compo- nent images themselves. In this paper we develop an opti- mal approach to recovering layer images and their associated motions from an arbitrary number of composite images. We develop two different techniques for estimating the compo- nent layer images given known motion estimates. The first approach uses constrained least squares to recover the layer images. The second approach iteratively refines lower and upper bounds on the layer images using two novel composit- ing operations, namely minimum- and maximum-composites of aligned images. We combine these layer extraction tech- niques with a dominant motion estimator and a subsequent motion refinement stage. This results in a completely auto- mated system that recovers transparent images and motions from a collection of input images.

This publication has 16 references indexed in Scilit: