Improving Deep Models of Person Re-identification for Cross-Dataset Usage - Artificial Intelligence Applications and Innovations (AIAI 2018) Access content directly
Conference Papers Year : 2018

Improving Deep Models of Person Re-identification for Cross-Dataset Usage

Abstract

Person re-identification (Re-ID) is the task of matching humans across cameras with non-overlapping views that has important applications in visual surveillance. Like other computer vision tasks, this task has gained much with the utilization of deep learning methods. However, existing solutions based on deep learning are usually trained and tested on samples taken from same datasets, while in practice one need to deploy Re-ID systems for new sets of cameras for which labeled data is unavailable. Here, we mitigate this problem for one state-of-the-art model, namely, metric embedding trained with the use of the triplet loss function, although our results can be extended to other models. The contribution of our work consists in developing a method of training the model on multiple datasets, and a method for its online practically unsupervised fine-tuning. These methods yield up to 19.1% improvement in Rank-1 score in the cross-dataset evaluation.
Fichier principal
Vignette du fichier
467708_1_En_7_Chapter.pdf (434.63 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01821073 , version 1 (22-06-2018)

Licence

Attribution

Identifiers

Cite

Sergey Rodionov, Alexey Potapov, Hugo Latapie, Enzo Fenoglio, Maxim Peterson. Improving Deep Models of Person Re-identification for Cross-Dataset Usage. 14th IFIP International Conference on Artificial Intelligence Applications and Innovations (AIAI), May 2018, Rhodes, Greece. pp.75-84, ⟨10.1007/978-3-319-92007-8_7⟩. ⟨hal-01821073⟩
130 View
19 Download

Altmetric

Share

Gmail Facebook X LinkedIn More