Construct Informative Triplet with Two-stage Hard-sample Generation

TL;DR

This paper introduces a two-stage generative approach for creating informative hard triplet samples, improving deep metric learning by enhancing sample quality and diversity through novel synthesis techniques.

Contribution

The paper presents a new two-stage hard sample generation framework using conditional GANs and adaptive constraints, advancing the quality and effectiveness of triplet-based learning.

Findings

Outperforms existing hard-sample generation algorithms on benchmark datasets.

Combining the proposed method with existing triplet mining strategies further improves performance.

Demonstrates robustness and superiority of the approach in deep metric learning tasks.

Abstract

In this paper, we propose a robust sample generation scheme to construct informative triplets. The proposed hard sample generation is a two-stage synthesis framework that produces hard samples through effective positive and negative sample generators in two stages, respectively. The first stage stretches the anchor-positive pairs with piecewise linear manipulation and enhances the quality of generated samples by skillfully designing a conditional generative adversarial network to lower the risk of mode collapse. The second stage utilizes an adaptive reverse metric constraint to generate the final hard samples. Extensive experiments on several benchmark datasets verify that our method achieves superior performance than the existing hard-sample generation algorithms. Besides, we also find that our proposed hard sample generation method combining the existing triplet mining strategies can further boost the deep metric learning performance.