Improving person re-identification performance by customized dataset and person detection

Research output: Contribution to conferencePaperAcademic

2 Citations (Scopus)
47 Downloads (Pure)

Abstract

For person re-identification (re-ID), nearly all person re-ID algorithms use public person re-ID datasets, where these datasets all consist of predefined image crops containing a single person. Unfortunately, these image crops are not optimal for video analysis, so that the person detection becomes suboptimal and person re-ID obtains a lower performance score. In this work, several techniques are presented that customize the person images of a popular public person re-ID dataset.
These techniques consist of customization algorithms based on postprocessing the person-detection bounding boxes using the original frames, resulting in several customized datasets to better facilitate person re-identification. We have evaluated five different ways for customization, based on widening the image crops, various aspect ratios and resolutions, and person instance segmentation. We have obtained a significant increase in performance with widened image crops, yielding a convincing performance increase of nearly 3% in the resulting Rank-1 score. Furthermore, when the applied random-cropping process is further optimized to this customization technique, an increase of even more than 4% is obtained. Both performance gains are a strong indication that any future person re-ID system may benefit from customizations based on the original video frames or from specializing the person detector.
Original languageEnglish
Pages268-1-268-9(9)
Number of pages8
DOIs
Publication statusPublished - 1 Feb 2019
EventIS&T International Symposium on Electronic Imaging 2019, Image Processing: Algorithms and Systems XVII - Burlingame, United States
Duration: 13 Jan 201917 Jan 2019
Conference number: XVII
http://www.imaging.org/site/IST/IST/Conferences/EI/EI_2019/Conference/C_IPAS.aspx

Conference

ConferenceIS&T International Symposium on Electronic Imaging 2019, Image Processing: Algorithms and Systems XVII
Abbreviated titleIPAS2019
CountryUnited States
CityBurlingame
Period13/01/1917/01/19
Internet address

Keywords

  • DukeMTMC
  • DukeMTMC-reID
  • Fixed aspect ratio
  • Image crop widening
  • Instance segmentation
  • Original camera output
  • Person detection
  • Person re-identification
  • Re-ID

Fingerprint

Dive into the research topics of 'Improving person re-identification performance by customized dataset and person detection'. Together they form a unique fingerprint.

Cite this