Cluster-Aware Similarity Diffusion for Instance Retrieval

TL;DR

This paper introduces Cluster-Aware Similarity diffusion, a novel method for instance retrieval that improves accuracy by restricting similarity propagation within local clusters, reducing misinformation from outliers.

Contribution

The paper proposes a new Cluster-Aware Similarity diffusion method with bidirectional diffusion and neighbor-guided smoothing to enhance instance retrieval accuracy.

Findings

Outperforms existing diffusion methods in retrieval tasks

Reduces influence of outliers and other manifolds

Validated on object re-identification benchmarks

Abstract

Diffusion-based re-ranking is a common method used for retrieving instances by performing similarity propagation in a nearest neighbor graph. However, existing techniques that construct the affinity graph based on pairwise instances can lead to the propagation of misinformation from outliers and other manifolds, resulting in inaccurate results. To overcome this issue, we propose a novel Cluster-Aware Similarity (CAS) diffusion for instance retrieval. The primary concept of CAS is to conduct similarity diffusion within local clusters, which can reduce the influence from other manifolds explicitly. To obtain a symmetrical and smooth similarity matrix, our Bidirectional Similarity Diffusion strategy introduces an inverse constraint term to the optimization objective of local cluster diffusion. Additionally, we have optimized a Neighbor-guided Similarity Smoothing approach to ensure similarity consistency among the local neighbors of each instance. Evaluations in instance retrieval and object re-identification validate the effectiveness of the proposed CAS, our code is publicly available.