Learning Deep Feature Representations with Domain Guided Dropout for Person Re-identification

Abstract

Learning generic and robust feature representations with data from multiple domains for the same problem is of great value, especially for the problems that have multiple datasets but none of them are large enough to provide abundant data variations. In this work, we present a pipeline for learning deep feature representations from multiple domains with Convolutional Neural Networks (CNNs). When training a CNN with data from all the domains, some neurons learn representations shared across several domains, while some others are effective only for a specific one. Based on this important observation, we propose a Domain Guided Dropout algorithm to improve the feature learning procedure. Experiments show the effectiveness of our pipeline and the proposed algorithm. Our methods on the person re-identification problem outperform state-of-the-art methods on multiple datasets by large margins.

arXiv commentTo appear in CVPR2016

References in corpus (12)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Improving neural networks by preventing co-adaptation of feature detectors
• Learning Transferable Features with Deep Adaptation Networks
• Unsupervised Domain Adaptation by Backpropagation
• Going Deeper with Convolutions
• Fully Convolutional Networks for Semantic Segmentation
• Rich feature hierarchies for accurate object detection and semantic segmentation
• Domain-Adversarial Neural Networks
• Person Re-identification Meets Image Search
• Learning to rank in person re-identification with metric ensembles
• Open-set Person Re-identification
• Factors of Transferability for a Generic ConvNet Representation

Cited by in corpus (14)

• Looking Beyond Appearances: Synthetic Training Data for Deep CNNs in Re-identification
• What-and-Where to Match: Deep Spatially Multiplicative Integration Networks for Person Re-identification
• Quality Aware Network for Set to Set Recognition
• Neural Person Search Machines
• Unsupervised Person Re-Identification: A Systematic Survey of Challenges and Solutions
• Frustratingly Easy Person Re-Identification: Generalizing Person Re-ID in Practice
• Distance-based Camera Network Topology Inference for Person Re-identification
• Learning Compact Appearance Representation for Video-based Person Re-Identification
• Cross Domain Knowledge Transfer for Person Re-identification
• Deep Feature Learning via Structured Graph Laplacian Embedding for Person Re-Identification
• Security Event Recognition for Visual Surveillance
• Learning Conditional Random Fields with Augmented Observations for Partially Observed Action Recognition
• Semantic Consistency and Identity Mapping Multi-Component Generative Adversarial Network for Person Re-Identification
• Learning Task-oriented Disentangled Representations for Unsupervised Domain Adaptation