Abstract:The number of existing person re-identification datasets is limited, and there are a series of changes such as illumination, background occlusion and pose among each dataset, which makes it difficult for the existing methods to learn robust feature representation, leading to a decline in recognition performance. To solve these problems, a person re-identification method combining style and pose generation is proposed in this paper. First with the impact of camera style differences in collecting images from different cameras, a style transformation method based on generative adversarial network is introduced into a person re-identification model, and cyclic generative adversarial networks (CycleGAN) is used to realize style transfer and reduce the influence of camera differences. Second, in view of the problem that when pedestrian pose changes greatly, easy to ignore identity-sensitive related information, AlphaPose is introduced to implement pose estimation. Combining style and posture for the first time and using improved deep convolution generative adversarial network (DCGAN) structure enrich the input sample information and generate unified style pose image; while using new synthetic data to train person re-identification network model improves the recognition performance of the model. Finally, further introducing random erasure method during data enhancement, in order to reduce the overfitting phenomena, improves the generalization ability of the network simultaneously and solves partial occlusion. The experimental results show that the proposed method outperforms typical style-based or pose-based methods. The accuracy of rank-1 and mAP on Market-1501 dataset is 90.4% and 74.5%, respectively, which are 2.28% and 5.78% higher, respectively. To a certain extent, the performance of person re-identification is improved.

Pose-Normalized Image Generation for Person Re-identification

Pose Invariant Person Re-Identification using Robust Pose-transformation GAN

A Unified Generative Adversarial Framework for Image Generation and Person Re-identification.

FD-GAN: Pose-guided Feature Distilling GAN for Robust Person Re-identification

Realistic Face Reenactment Via Self-Supervised Disentangling of Identity and Pose

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

Person Image Synthesis Through Siamese Generative Adversarial Network

Pose-guided Neural Network with Hybrid Representation for Person Re-Identification

Pose Variation Adaptation for Person Re- Identification.

Pose Transferrable Person Re-Identification

Adaptive Domain-Specific Normalization for Generalizable Person Re-Identification

Person Re-Identification Combined with Style Transfer and Pose Generation

Generate and Purify: Efficient Person Data Generation for Re-Identification

Crossing Generative Adversarial Networks for Cross-View Person Re-identification

Pose Transfer Generation with Semantic Parsing Attention Network for Person Re-Identification.

Pose and Color-Gamut Guided Generative Adversarial Network for Pedestrian Image Synthesis

Pose Attention-Guided Paired-Images Generation for Visible-Infrared Person Re-Identification

PGAN: Part-Based Nondirect Coupling Embedded GAN for Person Reidentification.

Pose-Guided Person Image Synthesis in the Non-Iconic Views.

Cross-dataset Person Re-Identification Using Similarity Preserved Generative Adversarial Networks.

CP-GAN: A Cross-Pose Profile Face Frontalization Boosting Pose-Invariant Face Recognition