CVMar 17, 2019

Robust Shape Regularity Criteria for Superpixel Evaluation

arXiv:1903.07146v311 citations
Originality Incremental advance
AI Analysis

This work addresses the need for better evaluation metrics in superpixel-based applications like object recognition or tracking, though it is incremental as it refines existing measurement approaches.

The authors tackled the problem of evaluating superpixel shape regularity by showing that existing circularity measures are inadequate, and they proposed a new metric based on convexity, balanced repartition, and contour smoothness, which is robust to scale and noise and improves method comparisons.

Regular decompositions are necessary for most superpixel-based object recognition or tracking applications. So far in the literature, the regularity or compactness of a superpixel shape is mainly measured by its circularity. In this work, we first demonstrate that such measure is not adapted for superpixel evaluation, since it does not directly express regularity but circular appearance. Then, we propose a new metric that considers several shape regularity aspects: convexity, balanced repartition, and contour smoothness. Finally, we demonstrate that our measure is robust to scale and noise and enables to more relevantly compare superpixel methods.

Foundations

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

Your Notes