Source-Guided Similarity Preservation for Online Person Re-Identification

Online Unsupervised Domain Adaptation (OUDA) for person Re-Identification (Re-ID) is the task of continuously adapting a model trained on a well-annotated source-domain dataset to a target domain observed as a data stream. In OUDA, person Re-ID models face two main challenges: catastrophic forgettin...

Full description

Saved in:
Bibliographic Details
Main Authors: Rami, Hamza, Giraldo, Jhony H., Winckler, Nicolas, Lathuiliere, Stephane
Format: Conference Proceeding
Language:English
Subjects:
accountable
Adaptation models
Algorithms
and algorithms
Benchmark testing
Computer vision
Data privacy
ethical computer vision
Explainable
fair
Feature extraction
formulations
Machine learning architectures
privacy-preserving
Regulation
Streaming media
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Online Unsupervised Domain Adaptation (OUDA) for person Re-Identification (Re-ID) is the task of continuously adapting a model trained on a well-annotated source-domain dataset to a target domain observed as a data stream. In OUDA, person Re-ID models face two main challenges: catastrophic forgetting and domain shift. In this work, we propose a new Source-guided Similarity Preservation (S2P) framework to alleviate these two problems. Our framework is based on the extraction of a support set composed of source images that maximizes the similarity with the target data. This support set is used to identify feature similarities that must be preserved during the learning process. S2P can incorporate multiple existing UDA methods to mitigate catastrophic forgetting. Our experiments show that S2P outperforms previous state-of-the-art methods on multiple real-to-real and synthetic-to-real challenging OUDA benchmarks.
ISSN:2642-9381
DOI:10.1109/WACV57701.2024.00173