Multi-Region bilinear convolutional neural networks for person re-identification

Abstract

In this work we propose a new architecture for person reidentification. As the task of re-identification is inherently associated with embedding learning and non-rigid appearance description, our architecture is based on the deep bilinear convolutional network (Bilinear-CNN) that has been proposed recently for fine-grained classification of highly non-rigid objects. While the last stages of the original Bilinear-CNN architecture completely removes the geometric information from consideration by performing orderless pooling, we observe that a better embedding can be learned by performing bilinear pooling in a more local way, where each pooling is confined to a predefined region. Our architecture thus represents a compromise between traditional convolutional networks and bilinear CNNs and strikes a balance between rigid matching and completely ignoring spatial information. We perform the experimental validation of the new architecture on the three popular benchmark datasets (Market-1501, CUHK01, CUHK03), comparing it to baselines that include Bilinear-CNN as well as prior art. The new architecture outperforms the baseline on all three datasets, while performing better than state-of-the-art on two out of three. The code and the pretrained models of the approach can be found at https://github.com/ madkn/MultiregionBilinearCNN-ReId.

Pyramidal Person Re-IDentification via Multi-Loss Dynamic Training

“…Part-based algorithms: By performing bilinear pooling in a more local way, an embedding can be learned, in which each pooling is confined to a predefined region [25]. Inspired by attention models, in [16,14,21], the attention-based deep neural networks are proposed to capture multiple attentions and select multi-scale attentive features.…”

Part-Aligned Bilinear Representations for Person Re-identification

“…This strategy has been widely adopted in fine-grained recogni- tion [53][54][55] and shows promising performance. For person re-identification, Ustinova et al [56] adopted a bilinear pooling to aggregate two different appearance maps; this method does not generate part-aligned representations and leads to poor performance. Our approach uses a bilinear pooling to aggregate appearance and part maps to compute part-aligned representations.…”

On Low-Resolution Face Recognition in the Wild: Comparisons and New Techniques

“…Wu et al [79] improved performance based on Ahmed's idea by using a deeper architecture and a new optimization method. Other deep network structures such as [69] and [65] have been designed which also effectively solved the ReID problem on older ReID datasets. Qui et al [54] attempted to perform facial ReID by using domain adaptation methods to reconcile different facial poses; however, their experiments were performed on the Multi-PIE [15] dataset, in which face images have controlled poses and illuminations.…”