LABNet: Local Graph Aggregation Network with Class Balanced Loss for Vehicle Re-Identification
This work provides an incremental improvement for vehicle re-identification, a computer vision task important for surveillance and autonomous driving applications.
This paper addresses vehicle re-identification by proposing a local graph aggregation network that learns associations of local information on the feature map, aiming to improve feature learning and reduce the effects of partial occlusion and background clutter. The method also introduces a class balanced loss to handle sample distribution imbalance in datasets, outperforming many state-of-the-art methods on three popular benchmarks.
Vehicle re-identification is an important computer vision task where the objective is to identify a specific vehicle among a set of vehicles seen at various viewpoints. Recent methods based on deep learning utilize a global average pooling layer after the backbone feature extractor, however, this ignores any spatial reasoning on the feature map. In this paper, we propose local graph aggregation on the backbone feature map, to learn associations of local information and hence improve feature learning as well as reduce the effects of partial occlusion and background clutter. Our local graph aggregation network considers spatial regions of the feature map as nodes and builds a local neighborhood graph that performs local feature aggregation before the global average pooling layer. We further utilize a batch normalization layer to improve the system effectiveness. Additionally, we introduce a class balanced loss to compensate for the imbalance in the sample distributions found in the most widely used vehicle re-identification datasets. Finally, we evaluate our method in three popular benchmarks and show that our approach outperforms many state-of-the-art methods.