2019 IEEE International Conference on Image Processing (ICIP) 2019
DOI: 10.1109/icip.2019.8803290
|View full text |Cite
|
Sign up to set email alerts
|

Deep Self-Paced Learning for Semi-Supervised Person Re-Identification Using Multi-View Self-Paced Clustering

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
2

Citation Types

0
24
0

Year Published

2020
2020
2021
2021

Publication Types

Select...
4
2
1

Relationship

0
7

Authors

Journals

citations
Cited by 14 publications
(24 citation statements)
references
References 12 publications
0
24
0
Order By: Relevance
“…Recently, semi‐supervised Re‐ID methods, such as dictionary learning, image matching, and metric learning, have garnered great attention. The self‐paced multi‐view clustering method 23 was proposed to generate pseudo labels for unlabeled training data. Two coherent dictionaries 24,25 related to the gallery and probe cameras were jointly learned from labeled and unlabeled images in order to bridge the variations in human appearance across cameras.…”
Section: Related Workmentioning
confidence: 99%
“…Recently, semi‐supervised Re‐ID methods, such as dictionary learning, image matching, and metric learning, have garnered great attention. The self‐paced multi‐view clustering method 23 was proposed to generate pseudo labels for unlabeled training data. Two coherent dictionaries 24,25 related to the gallery and probe cameras were jointly learned from labeled and unlabeled images in order to bridge the variations in human appearance across cameras.…”
Section: Related Workmentioning
confidence: 99%
“…To avoid the need for a large, labeled training dataset, Xin [28] proposed a self-paced multi-view clustering (SPMVC) method, which is a semi-supervised person Re-ID model trained with a small amount of labeled data and a large amount of unlabeled data. SPMVC performs the object Re-ID task using a heterogeneous set of CNNs initialized by the labeled training samples.…”
Section: Related Workmentioning
confidence: 99%
“…Guo et al used ranking loss and classification loss to match constraints under multiple feature scales and found that using small networks can also help to obtain much better search results [ 22 ]. In view of the fact that it is easy to obtain person data and difficult to obtain person labels in practical applications, researchers have tried to combine deep learning with semi-supervised learning algorithm [ 23 ] or unsupervised learning algorithm [ 24 ], and learn person features on unlabeled data through feature clustering and network fine-tuning strategy. Both semi-supervised and unsupervised algorithms have been often used in cross-domain person reID tasks for the purpose of improving the model’s generalization capability and mitigating the impact of domain gap, so as to finally increase the cross-domain prediction accuracy.…”
Section: Related Workmentioning
confidence: 99%