Boosting Weak Positives for Text Based Person Search

TL;DR

This paper introduces a boosting technique for text-based person search that dynamically emphasizes challenging positive samples during training, leading to improved retrieval performance across multiple datasets.

Contribution

It proposes a novel boosting-based method that dynamically weights weak positive pairs, addressing the issue of challenging samples being overlooked in existing models.

Findings

Improved performance on four pedestrian datasets.

Effectiveness of dynamic weighting of challenging samples.

Enhanced focus on hard positives improves retrieval accuracy.

Abstract

Large vision-language models have revolutionized cross-modal object retrieval, but text-based person search (TBPS) remains a challenging task due to limited data and fine-grained nature of the task. Existing methods primarily focus on aligning image-text pairs into a common representation space, often disregarding the fact that real world positive image-text pairs share a varied degree of similarity in between them. This leads models to prioritize easy pairs, and in some recent approaches, challenging samples are discarded as noise during training. In this work, we introduce a boosting technique that dynamically identifies and emphasizes these challenging samples during training. Our approach is motivated from classical boosting technique and dynamically updates the weights of the weak positives, wherein, the rank-1 match does not share the identity of the query. The weight allows these misranked pairs to contribute more towards the loss and the network has to pay more attention towards such samples. Our method achieves improved performance across four pedestrian datasets, demonstrating the effectiveness of our proposed module.