CVMar 17, 2019

SuperPatchMatch: an Algorithm for Robust Correspondences using Superpixel Patches

arXiv:1903.07169v239 citations
Originality Incremental advance
AI Analysis

This work addresses the challenge of unstable superpixel decompositions in computer vision, offering improved segmentation and labeling for applications like face and medical image analysis, though it is incremental as it builds on existing PatchMatch methods.

The paper tackles the problem of irregular superpixel segmentation by introducing SuperPatch, a superpixel-based patch structure that includes spatial information for robust descriptors, and SuperPatchMatch, a generalization of PatchMatch to these patches. The result outperforms state-of-the-art methods in computational cost and accuracy on face labeling and medical image segmentation tasks.

Superpixels have become very popular in many computer vision applications. Nevertheless, they remain underexploited since the superpixel decomposition may produce irregular and non stable segmentation results due to the dependency to the image content. In this paper, we first introduce a novel structure, a superpixel-based patch, called SuperPatch. The proposed structure, based on superpixel neighborhood, leads to a robust descriptor since spatial information is naturally included. The generalization of the PatchMatch method to SuperPatches, named SuperPatchMatch, is introduced. Finally, we propose a framework to perform fast segmentation and labeling from an image database, and demonstrate the potential of our approach since we outperform, in terms of computational cost and accuracy, the results of state-of-the-art methods on both face labeling and medical image segmentation.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes