Bilinear Attention Networks for Person Retrieval

By Pengfei Fang, Jieming Zhou, Soumava Kumar Roy, Lars Petersson, Mehrtash Harandi. 2019

156 citations recorded by the source corpus.

This paper investigates a novel Bilinear attention (Bi-attention) block, which discovers and uses second order statistical information in an input feature map, for the purpose of person retrieval. The Bi-attention block uses bilinear pooling to model the local pairwise feature interactions along each channel, while preserving the spatial structural information. We propose an Attention in Attention (AiA) mechanism to build inter-dependency among the second order local and global features with the intent to make better use of, or pay more attention to, such higher order statistical relationships. The proposed network, equipped with the proposed Bi-attention is referred to as Bilinear ATtention network (BAT-net). Our approach outperforms current state-of-the-art by a considerable margin across the standard benchmark datasets (e.g., CUHK03, Market-1501, DukeMTMC-reID and MSMT17).

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

On Learning the Geodesic Path for Incremental Learning (2021, methodology): “In this line of work, there strategy is to only update some parameters in which it is common using dropout [42] and attention [7, 18] for standard classification.”
Semantic-aware Knowledge Distillation for Few-Shot Class-Incremental Learning (2021, methodology): “We employ an attention module [8] to merge multiple embedding outputs and a loss to train the alignment appropriately with few-shot instances.”
Batch Coherence-Driven Network for Part-Aware Person Re-Identification (2020, unknown): “8 BAT-net [72] 95.”
Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification (2021, unknown): “9 BAT-Net [11] 95.”
Reinforced Attention for Few-Shot Learning and Beyond (2021, background): “The attention mechanism, aimed at attending to the discriminative areas adaptively, have been studied extensively for 2D and 3D visual tasks [17, 60, 56, 36, 69, 10, 39].”
Attention mechanisms in computer vision: A survey (2021, methodology): “[147] claimed that previous attention models only use first-order information and disregard higher-order statistical information.”
Hallucinating IDT Descriptors and I3D Optical Flow Features for Action Recognition With CNNs (2019, background): “Another aggregation family captures second- or higher-order statistics [8, 32, 37, 16].”
Dynamic Dual-Attentive Aggregation Learning for Visible-Infrared Person Re-Identification (2020, background): “mizing them separately. After concatenating the outputs of multiple heads, the graph structured attention-enhanced feature is then represented by xg i = ˚d X 8Ag(i;k)>0 g;l i;j h l(xo j )c L l=1 ; (8) and xg i is then robust to outlier samples, where ˚is the ELU activation function. To guide the ”
Relation-Aware Global Attention for Person Re-Identification (2019, background): “Recently, many studies resort to attention design to address the aforementioned challenges in person re-id by strengthening the discriminative features and suppressing interference [22, 39, 20, 12, 11, 6].”
Cross-Modality Person Re-Identification With Shared-Specific Feature Transfer (2020, background): “The recent works are mainly based on deep learning for more discriminative features [6, 18, 48, 55].”
Adaptive Subspaces for Few-Shot Learning (2020, background): “Other works use feature attention modules [28, 29] to modulate features for few-shot learning [30, 31].”