【PLPR】Progressive Learning for Person Re-Identification with One Example

Bibtex

@article{plpr,
title = {Progressive Learning for Person Re-Identification with One Example},
author = {Wu, Yu and Lin, Yutian and Dong, Xuanyi and Yan, Yan and Bian, Wei and Yang, Yi},
journal= {IEEE Transactions on Image Processing},
year = {2019},
volume = {28},
number = {6},
pages = {2872-2881},
doi = {10.1109/TIP.2019.2891895},
ISSN = {1057-7149},
month = {June},
}

Public information

IEEE Transactions on Image Processing (TIP), 2019

Fields

  • Person Re-ID
  • One-shot Learning

Code link

https://github.com/Yu-Wu/One-Example-Person-ReID

Main work

compare with EUG, author impoved the utilize of the unlabeled samples, aiming to optimize the accuracy of estimation for labels and selection for persudo sample with right labels.

Key technology

  • CNN
  • feature extraction
  • metric of samply similarity

Framework

【PLPR】Progressive Learning for Person Re-Identification with One Example

Dataset

  • Market-1501
  • DukeMTMC-reID
  • MARS
  • DukeMTMC-VideoReID

Results

【PLPR】Progressive Learning for Person Re-Identification with One Example
【PLPR】Progressive Learning for Person Re-Identification with One Example

Algorithm

【PLPR】Progressive Learning for Person Re-Identification with One Example

Others

论文下载: 关注下方公众号,回复“PLPR”即可获得论文原文

【PLPR】Progressive Learning for Person Re-Identification with One Example