Graph-Guided Sparse Reconstruction for Region Tagging

Abstract

Many of contextual correlations co-exist within the segmented regions among images, like the visual context and semantic context. The appropriate integration and utilization of such contexts are very important to boost the performance of region tagging. Inspired by the recent advances of sparse reconstruction methods, this paper proposes an approach, called Graph-Guided Sparse Reconstruction for Region Tagging (G <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> SRRT). The G <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> SRRT consists of two steps: sparse reconstruction for testing regions and tag propagation from training regions to testing regions. In G <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> SRRT, graph is conducted to flexibly model the contextual correlations among regions. To integrate the graph structure learned from training regions into the sparse reconstruction, we define a Graph-Guided Fusion (G <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> F) penalty over the graph to encourage the sparsity of differences between two reconstruction coefficients, which corresponds to the linked regions in the graph. Guided by this G <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> F penalty, the highly correlated regions tend to be jointly selected for the reconstruction, which results in a better performance of region tagging. Experiments on three open benchmark image datasets demonstrate the effectiveness of the proposed algorithm.