Adversarial Pairwise Reverse Attention for Camera Performance Imbalance in Person Re-identification: New Dataset and Metrics
Authors:
Eugene P. W. Ang,
Shan Lin,
Rahul Ahuja,
Nemath Ahmed,
Alex C. Kot
Abstract:
Existing evaluation metrics for Person Re-Identification (Person ReID) models focus on system-wide performance. However, our studies reveal weaknesses due to the uneven data distributions among cameras and different camera properties that expose the ReID system to exploitation. In this work, we raise the long-ignored ReID problem of camera performance imbalance and collect a real-world privacy-awa…
▽ More
Existing evaluation metrics for Person Re-Identification (Person ReID) models focus on system-wide performance. However, our studies reveal weaknesses due to the uneven data distributions among cameras and different camera properties that expose the ReID system to exploitation. In this work, we raise the long-ignored ReID problem of camera performance imbalance and collect a real-world privacy-aware dataset from 38 cameras to assist the study of the imbalance issue. We propose new metrics to quantify camera performance imbalance and further propose the Adversarial Pairwise Reverse Attention (APRA) Module to guide the model learning the camera invariant feature with a novel pairwise attention inversion mechanism.
△ Less
Submitted 4 July, 2022;
originally announced July 2022.
DEX: Domain Embedding Expansion for Generalized Person Re-identification
Authors:
Eugene P. W. Ang,
Lin Shan,
Alex C. Kot
Abstract:
In recent years, supervised Person Re-identification (Person ReID) approaches have demonstrated excellent performance. However, when these methods are applied to inputs from a different camera network, they typically suffer from significant performance degradation. Different from most domain adaptation (DA) approaches addressing this issue, we focus on develo** a domain generalization (DG) Perso…
▽ More
In recent years, supervised Person Re-identification (Person ReID) approaches have demonstrated excellent performance. However, when these methods are applied to inputs from a different camera network, they typically suffer from significant performance degradation. Different from most domain adaptation (DA) approaches addressing this issue, we focus on develo** a domain generalization (DG) Person ReID model that can be deployed without additional fine-tuning or adaptation. In this paper, we propose the Domain Embedding Expansion (DEX) module. DEX dynamically manipulates and augments deep features based on person and domain labels during training, significantly improving the generalization capability and robustness of Person ReID models to unseen domains. We also developed a light version of DEX (DEXLite), applying negative sampling techniques to scale to larger datasets and reduce memory usage for multi-branch networks. Our proposed DEX and DEXLite can be combined with many existing methods, Bag-of-Tricks (BagTricks), the Multi-Granularity Network (MGN), and Part-Based Convolutional Baseline (PCB), in a plug-and-play manner. With DEX and DEXLite, existing methods can gain significant improvements when tested on other unseen datasets, thereby demonstrating the general applicability of our method. Our solution outperforms the state-of-the-art DG Person ReID methods in all large-scale benchmarks as well as in most the small-scale benchmarks.
△ Less
Submitted 21 October, 2021;
originally announced October 2021.
