Conference paper · Book chapter
Global Similarity with Additive Smoothness for Spectral Segmentation
Visual Computing, Department of Applied Mathematics and Computer Science, Technical University of Denmark1
Statistics and Data Analysis, Department of Applied Mathematics and Computer Science, Technical University of Denmark2
Department of Applied Mathematics and Computer Science, Technical University of Denmark3
Faithful representation of pairwise pixel affinities is crucial for the outcome of spectral segmentation methods. In conventional affinity models only close-range pixels interact, and a variety of subsequent techniques aims at faster propagation of local grouping cues across longrange connections. In this paper we propose a general framework for constructing a full-range affinity matrix.
Our affinity matrix consists of a global similarity matrix and an additive proximity matrix. The similarity in appearance, including intensity and texture, is encoded for each pair of image pixels. Despite being full-range, our similarity matrix has a simple decomposition, which exploits an assignment of image pixels to dictionary elements.
The additive proximity enforces smoothness to the segmentation by imposing interactions between near-by pixels. Our approach allows us to assess the advantages of using a full-range affinity for various spectral segmentation problems. Within our general framework we develop a few variants of full affinity for experimental validation.
The performance we accomplish on composite textured images is excellent, and the results on natural images are promising.
Language: | English |
---|---|
Publisher: | Springer |
Year: | 2019 |
Pages: | 357-368 |
Proceedings: | 7th International Conference on Scale Space and Variational Methods in Computer Vision |
Series: | Lecture Notes in Computer Science |
Journal subtitle: | 7th International Conference, Ssvm 2019, Hofgeismar, Germany, June 30 – July 4, 2019, Proceedings |
ISBN: | 3030223671 , 303022368X , 303022368x , 9783030223670 and 9783030223687 |
ISSN: | 03029743 and 16113349 |
Types: | Conference paper and Book chapter |
DOI: | 10.1007/978-3-030-22368-7_28 |
ORCIDs: | Dahl, Vedrana Andersen , Dahl, Anders Bjorholm and 0000-0002-5243-0331 |