Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
1 file | 2.36 MB
Add to list
  1. Search results
  2. Resolution based feature distillation...

Resolution based feature distillation for cross resolution person re-identification

Asad Munir, Chengjin Lyu (UGent) , Bart Goossens (UGent) , Wilfried Philips (UGent) and Christian Micheloni
(2021) 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). In IEEE International Conference on Computer Vision Workshops p.281-289
Author
Organization
Project
  • ACHIEVE (AdvanCed Hardware/Software Components for Integrated/Embedded Vision SystEms (ACHIEVE))
Abstract
Person re-identification (re-id) aims to retrieve images of same identities across different camera views. Resolution mismatch occurs due to varying distances between person of interest and cameras, this significantly degrades the performance of re-id in real world scenarios. Most of the existing approaches resolve the re-id task as low resolution problem in which a low resolution query image is searched in a high resolution images gallery. Several approaches apply image super resolution techniques to produce high resolution images but ignore the multiple resolutions of gallery images which is a better realistic scenario. In this paper, we introduce channel correlations to improve the learning of features from the degraded data. In addition, to overcome the problem of multiple resolutions we propose a Resolution based Feature Distillation (RFD) approach. Such an approach learns resolution invariant features by filtering the resolution related features from the final feature vectors that are used to compute the distance matrix. We tested the proposed approach on two synthetically created datasets and on one original multi resolution dataset with real degradation. Our approach improves the performance when multiple resolutions occur in the gallery and have comparable results in case of single resolution (low resolution re-id).

Downloads

  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 2.36 MB

Citation

Please use this url to cite or link to this publication:

MLA
Munir, Asad, et al. “Resolution Based Feature Distillation for Cross Resolution Person Re-Identification.” 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), IEEE, 2021, pp. 281–89, doi:10.1109/iccvw54120.2021.00036.
APA
Munir, A., Lyu, C., Goossens, B., Philips, W., & Micheloni, C. (2021). Resolution based feature distillation for cross resolution person re-identification. 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 281–289. https://doi.org/10.1109/iccvw54120.2021.00036
Chicago author-date
Munir, Asad, Chengjin Lyu, Bart Goossens, Wilfried Philips, and Christian Micheloni. 2021. “Resolution Based Feature Distillation for Cross Resolution Person Re-Identification.” In 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 281–89. IEEE. https://doi.org/10.1109/iccvw54120.2021.00036.
Chicago author-date (all authors)
Munir, Asad, Chengjin Lyu, Bart Goossens, Wilfried Philips, and Christian Micheloni. 2021. “Resolution Based Feature Distillation for Cross Resolution Person Re-Identification.” In 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 281–289. IEEE. doi:10.1109/iccvw54120.2021.00036.
Vancouver
1.
Munir A, Lyu C, Goossens B, Philips W, Micheloni C. Resolution based feature distillation for cross resolution person re-identification. In: 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW). IEEE; 2021. p. 281–9.
IEEE
[1]
A. Munir, C. Lyu, B. Goossens, W. Philips, and C. Micheloni, “Resolution based feature distillation for cross resolution person re-identification,” in 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), Montreal, BC, Canada, 2021, pp. 281–289.
@inproceedings{8733608,
  abstract     = {{Person re-identification (re-id) aims to retrieve images of same identities across different camera views. Resolution mismatch occurs due to varying distances between person of interest and cameras, this significantly degrades the performance of re-id in real world scenarios. Most of the existing approaches resolve the re-id task as low resolution problem in which a low resolution query image is searched in a high resolution images gallery. Several approaches apply image super resolution techniques to produce high resolution images but ignore the multiple resolutions of gallery images which is a better realistic scenario. In this paper, we introduce channel correlations to improve the learning of features from the degraded data. In addition, to overcome the problem of multiple resolutions we propose a Resolution based Feature Distillation (RFD) approach. Such an approach learns resolution invariant features by filtering the resolution related features from the final feature vectors that are used to compute the distance matrix. We tested the proposed approach on two synthetically created datasets and on one original multi resolution dataset with real degradation. Our approach improves the performance when multiple resolutions occur in the gallery and have comparable results in case of single resolution (low resolution re-id).}},
  author       = {{Munir, Asad and Lyu, Chengjin and Goossens, Bart and Philips, Wilfried and Micheloni, Christian}},
  booktitle    = {{2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW)}},
  isbn         = {{9781665401913}},
  issn         = {{2473-9936}},
  language     = {{eng}},
  location     = {{Montreal, BC, Canada}},
  pages        = {{281--289}},
  publisher    = {{IEEE}},
  title        = {{Resolution based feature distillation for cross resolution person re-identification}},
  url          = {{http://doi.org/10.1109/iccvw54120.2021.00036}},
  year         = {{2021}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: