A graph-cut approach to image segmentation using an affinity graph based on l0−sparse representation of features

Abstract : We propose a graph-cut based image segmentation method by constructing an affinity graph using l0 sparse representation. Computing first oversegmented images, we associate with all segments, that we call superpixels, a collection of features. We find the sparse representation of each set of features over the dictionary of all features by solving a l0-minimization problem. Then, the connection information between superpixels is encoded as the non-zero representation coefficients, and the affinity of connected superpixels is derived by the corresponding representation error. This provides a l0 affinity graph that has interesting properties of long range and sparsity, and a suitable graph cut yields a segmentation. Experimental results on the BSD database demonstrate that our method provides perfectly semantic regions even with a constant segmentation number, but also that very competitive quantitative results are achieved.
Complete list of metadatas

Cited literature [22 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-00833278
Contributor : Simon Masnou <>
Submitted on : Friday, July 26, 2013 - 1:43:37 PM
Last modification on : Thursday, November 21, 2019 - 2:25:30 AM
Long-term archiving on: Sunday, October 27, 2013 - 2:55:10 AM

File

SR-hal.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00833278, version 1

Citation

Xiaofang Wang, Huibin Li, Charles-Edmond Bichot, Simon Masnou, Liming Chen. A graph-cut approach to image segmentation using an affinity graph based on l0−sparse representation of features. IEEE International Conference on Image Processing, Sep 2013, Melbourne, Australia. pp.4019-4023. ⟨hal-00833278⟩

Share

Metrics

Record views

687

Files downloads

1375