Deep Representation Learning With Part Loss for Person Re-Identification

Abstract

Learning discriminative representations for unseen person images is critical for person Re-Identification (ReID). Most of current approaches learn deep representations in classification tasks, which essentially minimize the empirical classification risk on the training set. As shown in our experiments, such representations easily get overfitted on a discriminative human body part among the training set. To gain the discriminative power on unseen person images, we propose a deep representation learning procedure named Part Loss Networks (PL-Net), to minimize both the empirical classification risk and the representation learning risk. The representation learning risk is evaluated by the proposed part loss, which automatically detects human body parts, and computes the person classification loss on each part separately. Compared with traditional global classification loss, simultaneously considering part loss enforces the deep network to learn representations for different parts and gain the discriminative power on unseen persons. Experimental results on three person ReID datasets, i.e., Mar-ket1501, CUHK03, VIPeR, show that our representation outperforms existing deep representations.

ABD-Net: Attentive but Diverse Person Re-Identification

“…Zhao et al [8] proposed a part-aligned representation based on a part map detector for each predefined body part. Yao et al [9] proposed a Part Loss Network which defined a loss for each average pooled body part and jointly optimized the summation losses. Si et al [10] proposed a dual attention matching network based on an inter-class and an intra-class attention module to capture the context information of video sequences for person Re-ID.…”

“…Our proposed attention mechanism differs from previous methods in several aspects. First, previous methods [8,9,11] only use attention mechanisms to extract partbased spatial patterns from person images, which are usually focus in the foregrounds. In contrast, ABD-Net combines spatial and channel clues; besides, our added diversity constraint will avoid the overly correlated and redundant attentive features.…”

“…Motivated by such observations, the attention mechanism [7] was introduced to enforce the features to mainly capture the discriminative appearances of human bodies (or certain body parts). Since then, the attention-based models [8,9,10,11,12] have boosted person Re-ID performance much.…”

AANet: Attribute Attention Network for Person Re-Identifications

“…The local body parts are detected and learned together with the global image by the four-stream CNN model, which yields a discriminatory and robust representation. Yao et al [26] proposed the Part Loss Networks (PL-Net) to automatically detect human parts and cross train them with the main identity task. Zhao et al [27] follows the concept of attention model and uses a part map detector to extract multiple body regions in order to compute their corresponding representations.…”

Learning Resolution-Invariant Deep Representations for Person Re-Identification

“…Related studies are performed on this problem. For pedestrian similarity issue, [1], [2] adopted a deep convolutional network and body parts to learn discriminative representations. They rely on powerful CNN to extract feature representations.…”