# Learning Incremental Triplet Margin for Person Re-identification

arXiv: https://arxiv.org/abs/1812.06576 · HTML: https://mlsanity.com/paper/1812.06576/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 17 Dec 2018.

## Built on by (most cited first, 9 of 9)

- [Transformer Decoders with MultiModal Regularization for Cross-Modal Food Retrieval](https://mlsanity.com/paper/2204.09730/) (2022)
- [Relevance-based Margin for Contrastively-trained Video Retrieval Models](https://mlsanity.com/paper/2204.13001/) (2022)
- [Improving Video Retrieval by Adaptive Margin](https://mlsanity.com/paper/2303.05093/) (2023)
- [Deep Point Cloud Normal Estimation via Triplet Learning](https://mlsanity.com/paper/2110.10494/) (2021)
- [FPB: Feature Pyramid Branch for Person Re-Identification](https://mlsanity.com/paper/2108.01901/) (2021)
- [DeepSSN: a deep convolutional neural network to assess spatial scene similarity](https://mlsanity.com/paper/2202.04755/) (2022)
- [POA: Pre-training Once for Models of All Sizes](https://mlsanity.com/paper/2408.01031/) (2024)
- [RankingMatch: Delving into Semi-Supervised Learning with Consistency Regularization and Ranking Loss](https://mlsanity.com/paper/2110.04430/) (2021)
- [Maintaining Difficulty: A Margin Scheduler for Triplet Loss in Siamese Networks Training](https://mlsanity.com/paper/2603.26389/) (2026)

## Abstract (arXiv)

Person re-identification (ReID) aims to match people across multiple non-overlapping video cameras deployed at different locations. To address this challenging problem, many metric learning approaches have been proposed, among which triplet loss is one of the state-of-the-arts. In this work, we explore the margin between positive and negative pairs of triplets and prove that large margin is beneficial. In particular, we propose a novel multi-stage training strategy which learns incremental triplet margin and improves triplet loss effectively. Multiple levels of feature maps are exploited to make the learned features more discriminative. Besides, we introduce global hard identity searching method to sample hard identities when generating a training batch. Extensive experiments on Market-1501, CUHK03, and DukeMTMCreID show that our approach yields a performance boost and outperforms most existing state-of-the-art methods.