PAC-GAN: An Effective Pose Augmentation Scheme for Unsupervised Cross-View Person Re-identification

Keyphrases

• Person Re-identification 100%
• Cross-view 100%
• Re-ID 100%
• Pose Augmentation 100%
• Recognition Rate 40%
• Unsupervised Method 20%
• Pedestrian 20%
• Deep Learning Methods 20%
• Network Approach 20%
• Generation Model 20%
• Image Retrieval 20%
• Human Motion 20%
• Existing Data 20%
• Body Posture 20%
• Original Image 20%
• Disjoint 20%
• Shared Strategy 20%
• Formal Definition 20%
• Weight Sharing 20%
• Non-overlapping Cameras 20%
• Pose Generation 20%
• Retrieval Benchmark 20%
• Deep Model 20%
• Generated Samples 20%
• Pedestrian Image 20%
• Conditional Generative Adversarial Network (cGAN) 20%

Computer Science

• Recognition Rate 100%
• Person Re-Identification 100%
• Priori Knowledge 50%
• Generation Model 50%
• Image Retrieval 50%
• Unsupervised Manner 50%
• Deep Learning 50%
• Human Motions 50%
• Formal Definition 50%
• Generative Adversarial Networks 50%