Re-ID done right: towards good practices for person re-identification

Abstract

Training a deep architecture using a ranking loss has become standard for the person re-identification task. Increasingly, these deep architectures include additional components that leverage part detections, attribute predictions, pose estimators and other auxiliary information, in order to more effectively localize and align discriminative image regions. In this paper we adopt a different approach and carefully design each component of a simple deep architecture and, critically, the strategy for training it effectively for person re-identification. We extensively evaluate each design choice, leading to a list of good practices for person re-identification. By following these practices, our approach outperforms the state of the art, including more complex methods with auxiliary components, by large margins on four benchmark datasets. We also provide a qualitative analysis of our trained representation which indicates that, while compact, it is able to capture information from localized and discriminative regions, in a manner akin to an implicit attention mechanism.

References in corpus (12)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Deep Residual Learning for Image Recognition
• Densely Connected Convolutional Networks
• Pedestrian Alignment Network for Large-scale Person Re-identification
• Rethinking the Inception Architecture for Computer Vision
• Aggregated Residual Transformations for Deep Neural Networks
• SVDNet for Pedestrian Retrieval
• HydraPlus-Net: Attentive Deep Features for Pedestrian Analysis
• Deep Image Retrieval: Learning global representations for image search
• Deep Mutual Learning
• Smart Mining for Deep Metric Learning
• Neural Person Search Machines

Cited by in corpus (23)

• Stripe-based and Attribute-aware Network: A Two-Branch Deep Model for Vehicle Re-identification
• Relation-Aware Global Attention for Person Re-identification
• AANet: Attribute Attention Network for Person Re-Identifications
• Batch DropBlock Network for Person Re-identification and Beyond
• Style Normalization and Restitution for Generalizable Person Re-identification
• Densely Semantically Aligned Person Re-Identification
• Joint Detection and Tracking in Videos with Identification Features
• Multi-Granularity Reference-Aided Attentive Feature Aggregation for Video-based Person Re-identification
• Semantics-Aligned Representation Learning for Person Re-identification
• Improving Deep Visual Representation for Person Re-identification by Global and Local Image-language Association
• Taking Modality-free Human Identification as Zero-shot Learning
• Global Distance-distributions Separation for Unsupervised Person Re-identification
• Rethinking the Distribution Gap of Person Re-identification with Camera-based Batch Normalization
• Apparel-invariant Feature Learning for Apparel-changed Person Re-identification
• Learning to Learn in a Semi-Supervised Fashion
• Hybrid-Attention Guided Network with Multiple Resolution Features for Person Re-Identification
• Diversity-Achieving Slow-DropBlock Network for Person Re-Identification
• HorNet: A Hierarchical Offshoot Recurrent Network for Improving Person Re-ID via Image Captioning
• Progressive Learning Algorithm for Efficient Person Re-Identification
• End-to-End Training of CNN Ensembles for Person Re-Identification
• SCPNet: Spatial-Channel Parallelism Network for Joint Holistic and Partial Person Re-Identification
• Sparse Label Smoothing Regularization for Person Re-Identification
• Beyond Triplet Loss: Meta Prototypical N-tuple Loss for Person Re-identification