%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2011/07.09.02.10
%2 sid.inpe.br/sibgrapi/2011/07.09.02.10.45
%@doi 10.1109/SIBGRAPI.2011.34
%T Spectral Segmentation using Cartoon-Texture Decomposition and Inner Product-based metric
%D 2011
%A Casaca, Wallace,
%A Paiva, Afonso,
%A Nonato, Luis Gustavo,
%@affiliation Instituto de Ciências Matemáticas e de Computação – USP
%@affiliation Instituto de Ciências Matemáticas e de Computação – USP
%@affiliation Instituto de Ciências Matemáticas e de Computação – USP
%E Lewiner, Thomas,
%E Torres, Ricardo,
%B Conference on Graphics, Patterns and Images, 24 (SIBGRAPI)
%C Maceió, AL, Brazil
%8 28-31 Aug. 2011
%I IEEE Computer Society
%J Los Alamitos
%S Proceedings
%K Spectral cut, image segmentation, similarity graph, cartoon-texture decomposition, harmonic analysis, normalized cuts.
%X This paper presents a user-assisted image partition technique that combines cartoon-texture decomposition, inner product-based similarity metric, and spectral cut into a unified framework. The cartoon-texture decomposition is used to first split the image into textured and texture-free components, the latter being used to define a gradient-based inner-product function. An affinity graph is then derived and weights are assigned to its edges according to the inner product-based metric. Spectral cut is computed on the affinity graph so as to partition the image. The computational burden of the spectral cut is mitigated by a fine-to-coarse image representation process, which enables moderate size graphs that can be handled more efficiently. The partitioning can be steered by interactively by changing the weights of the graph through user strokes. Weights are updated by combining the texture component computed in the first stage of our pipeline and a recent harmonic analysis technique that captures waving patterns. Finally, a coarse-to-fine interpolation is applied in order to project the partition back onto the original image. The suitable performance of the proposed methodology is attested by comparisons against state-of-art spectral segmentation methods.
%@language en
%3 Sibgrapi_paper.pdf