Unsupervised person re-identification via local manifold consistency learning

Zhou, Ruiling; He, Rui

doi:10.1117/1.jei.31.3.033017

Cited by 1 publication

(8 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Notably, our proposed method exhibits superior performance on all three datasets compared to other camera-aware learning methods. For example, our method obtains 1.4% Rank-1 and 2.9% mAP increase on Market-1501, and 0.7% Rank-1 and 0.9% mAP increase on DukeMTMC compared with the second-best camera-aware approach LMCL 7 . The main differences between LMCL 7 and our JGPP exist in the approach of exploiting comprehensive features.…”

Section: Methodsmentioning

confidence: 89%

“…For example, our method obtains 1.4% Rank-1 and 2.9% mAP increase on Market-1501, and 0.7% Rank-1 and 0.9% mAP increase on DukeMTMC compared with the second-best camera-aware approach LMCL. 7 The main differences between LMCL 7 and our JGPP exist in the approach of exploiting comprehensive features. LMCL 7 crops the image into different regions randomly and selectively focuses on more relevant regions with an attention mechanism.…”

Section: Comparison Resultsmentioning

confidence: 99%

“…Afterward, the camera-aware methods used for comparison include SSL, 23 JVTC, 40 IICS, 24 CAP, 26 IIDS, 25 LMCL, 7 and MHOR 41 . SSL 23 proposes a re-ID framework with re-assigned soft label distribution learned from similar images, and it also employs camera labels and partial details of each sample image to guide similarity estimation.…”

Section: Methodsmentioning

confidence: 99%

“…CAP 26 splits each cluster into multiple proxies according to different cameras and generates more reliable pseudo labels for the intra-camera and inter-camera contrastive learning. LMCL 7 is a local manifold consistency learning framework that consists of a context-aware feature embedding network and a camera-aware manifold alignment strategy. MHOR 41 presents a transformer-based network that includes the intra-camera transformer and inter-camera transformer.…”

Section: Methodsmentioning

confidence: 99%

“…The main differences between LMCL 7 and our JGPP exist in the approach of exploiting comprehensive features. LMCL 7 crops the image into different regions randomly and selectively focuses on more relevant regions with an attention mechanism. Hence, there is a lack of feature stability in LMCL 7 .…”

Section: Methodsmentioning

confidence: 99%

See 4 more Smart Citations

Joint global feature and part-based pyramid features for unsupervised person re-identification

Zhang,

Fan,

Zhou

et al. 2024

J. Electron. Imag.

View full text Add to dashboard Cite

In recent years, unsupervised person re-identification (re-ID) has attracted a lot of attention because it can save manual annotation costs and adapt to new scenes more conveniently in real-world applications. We focus on fully unsupervised learning-based re-ID, which aims to train a discriminative model based on unlabeled person images. In unsupervised learning, we need to generate pseudo labels by clustering convolutional features and then train convolutional neural network (CNN) models with these pseudo labels. The features used in the clustering process play an important role to ensure the quality of pseudo labels. Hence, we propose an enhancing feature extraction method to increase the reliability of generated pseudo labels and thus facilitate CNN model training. In order to enrich the obtained features, we carry out the feature extraction from both global and local aspects. The global features are extracted with ResNet50 backbone as many existing methods do. The local features are extracted by an additional part-based feature pyramid structure, in which the person image is divided into three parts and the features are extracted from each part with a pyramid structure. Then, we fuse the multi-layer pyramid features for each part as the local features. According to the joint global features and local features, the pseudo labels are predicted using clustering algorithms and further refined based on the similarity between global and local features. In addition, we design an inter-camera association learning component to effectively learn the ID discrimination ability across cameras. Extensive experiments on three large and representative person re-ID datasets demonstrate the effectiveness of the proposed approach.

show abstract

Section: Methodsmentioning

confidence: 89%

Section: Comparison Resultsmentioning

confidence: 99%