Vehicle re-identification (Re-ID) distinguishes between the same vehicle and other vehicles in images. It is challenging due to significant intra-instance differences between identical vehicles from different views and subtle inter-instance differences of similar vehicles. Researchers have tried to address this problem by extracting features robust to variations of viewpoints and environments and using additional metadata such as key points, orientation, and temporal information, which require expensive annotations. More recently, many efforts have been made to improve performance by adopting the attention mechanism. In this paper, we propose a multiple soft attention network to provide part-aware attention weights and extract more representative and robust features for vehicle Re-ID. This model uses multiple spatial and channel-wise attention branches to allow soft partitioning of images and extracting more representative and robust features for vehicle Re-Id.Our approach achieved state-of-the-art (SOTA) performance among the approaches that did not use metadata and showed comparable performance to the methods that used metadata in the experiments using the VehicleID and VeRi-776 datasets.
Multiple Soft Attention Network for Vehicle Re-Identification
08.10.2022
1462123 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Vehicle identification statistical method based on multi-attention mechanism network
Europäisches Patentamt | 2023
|Source-attention features fusion network for person re-identification
British Library Conference Proceedings | 2022
|A Novel Dual-pooling Attention Module for UAV Vehicle Re-identification
ArXiv | 2023
|