Image Matting with KL-Divergence Based Sparse Sampling

IEEE International Conference on Computer Vision (ICCV) 2015

Levent Karacan     Aykut Erdem     Erkut Erdem

Input
Robust
Shared
Global
Comprehensive
Proposed

We use a recent sparse subset selection algorithm to obtain color samples using a dissimilarity matrix for which we propose a KL-Divergence based statistical measure. Proposed distance measure provides better discrimination than color distance. Our image matting method collects representative color samples for unknown region from all known background and foreground regions without making any spatial assumption in contrast to previous sampling based image matting methods. As a result, our approach provides highly competitive results against the state-of-the-art methods.

Abstract

Previous sampling-based image matting methods typically rely on certain heuristics in collecting representative samples from known regions, and thus their performance deteriorates if the underlying assumptions are not satisfied. To alleviate this, in this paper we take an entirely new approach and formulate sampling as a sparse subset selection problem where we propose to pick a small set of candidate samples that best explains the unknown pixels. Moreover, we describe a new distance measure for comparing two samples which is based on KL-divergence between the distributions of features extracted in the vicinity of the samples. Using a standard benchmark dataset for image matting, we demonstrate that our approach provides more accurate results compared with the state-of-the-art methods.

Paper

L.Karacan,A. Erdem and E. Erdem. Image Matting with KL-Divergence Based Sparse Sampling, in IEEE International Conference on Computer Vision (ICCV). 2015.

Code

Matlab Code [.zip]

Supplementary Material

Supplemental [.pdf]

Acknowledgments

This research was supported in part by The Scientific and Technological Research Council of Turkey (TUBITAK), Career Development Award 112E146.