Interpretable And Generalizable Person Re-identification With Query-adaptive Convolution And Temporal Lifting

Abstract

For person re-identification, existing deep networks often focus on representation learning. However, without transfer learning, the learned model is fixed as is, which is not adaptable for handling various unseen scenarios. In this paper, beyond representation learning, we consider how to formulate person image matching directly in deep feature maps. We treat image matching as finding local correspondences in feature maps, and construct query-adaptive convolution kernels on the fly to achieve local matching. In this way, the matching process and results are interpretable, and this explicit matching is more generalizable than representation features to unseen scenarios, such as unknown misalignments, pose or viewpoint changes. To facilitate end-to-end training of this architecture, we further build a class memory module to cache feature maps of the most recent samples of each class, so as to compute image matching losses for metric learning. Through direct cross-dataset evaluation, the

Stats

• citations125
• S2 citations—
• github stars218
• HF likes0
• heat score20.43
• arxiv keyliao2019interpretable

Related papers

• Structured Deep Hashing With Convolutional Neural Networks For Fast Person Re-identification (2017)12.87
• Cross-entropy Adversarial View Adaptation For Person Re-identification (2019)12.93
• Transmatcher: Deep Image Matching Through Transformers For Generalizable Person Re-identification (2021)4.68
• A Pose-sensitive Embedding For Person Re-identification With Expanded Cross Neighborhood Re-ranking (2017)23.25
• A Convolutional Baseline For Person Re-identification Using Vision And Language Descriptions (2020)0.00
• Deep-person: Learning Discriminative Deep Features For Person Re-identification (2017)16.90
• Deep Co-attention Based Comparators For Relative Representation Learning In Person Re-identification (2018)13.34
• Part-aligned Bilinear Representations For Person Re-identification (2018)19.30