Embedding Deep Metric for Person Re-identification: A Study Against Large Variations

Abstract

Person re-identification is challenging due to the large variations of pose, illumination, occlusion and camera view. Owing to these variations, the pedestrian data is distributed as highly-curved manifolds in the feature space, despite the current convolutional neural networks (CNN)'s capability of feature extraction. However, the distribution is unknown, so it is difficult to use the geodesic distance when comparing two samples. In practice, the current deep embedding methods use the Euclidean distance for the training and test. On the other hand, the manifold learning methods suggest to use the Euclidean distance in the local range, combining with the graphical relationship between samples, for approximating the geodesic distance. From this point of view, selecting suitable positive (i.e. intra-class) training samples within a local range is critical for training the CNN embedding, especially when the data has large intra-class variations. In this paper, we propose a novel moderate positive sample mining method to train robust CNN for person re-identification, dealing with the problem of large variation. In addition, we improve the learning by a metric weight constraint, so that the learned metric has a better generalization ability. Experiments show that these two strategies are effective in learning robust deep metrics for person re-identification, and accordingly our deep model significantly outperforms the state-of-the-art methods on several benchmarks of person re-identification. Therefore, the study presented in this paper may be useful in inspiring new designs of deep models for person re-identification.

Impostor Resilient Multimodal Metric Learning for Person Reidentification

“…These global metrics [16][17][18][19] project features into low dimension subspace where they tend to maximize the discrimination among different persons; however, these metrics still suffer a great challenge from impostor (an impostor is a person that belongs to the other person and, however, possess higher similarity with the given query than the right Gallery sample) samples [20,21]. Though, in past some attempts are made to eliminate impostors [14,[20][21][22], however, all these attempts have not given due consideration of different transform modals on which the reidentification images lie [23].…”

Deep Spatial-Semantic Attention for Fine-Grained Sketch-Based Image Retrieval

“…Compared to first-order alternatives, our energy function is more robust against misalignment between sketch and photo channels, and can accommodate better the more detailed but noisier fine-grained feature map representation. Mahalanobis distance [41,35] is another example of a higher-order energy function in that it does O(N 2 ) comparisons for N channels. However it is based on elementwise difference followed by bilinear product so the effect is to learn which dimension pairs are important to match, rather than compensate for misalignment and noise between the input vectors.…”

“…They have been used mainly for multi-view fusion, for example, fusing the text and image embeddings in visual question answering [21] and zero-shot recognition [4]. Outer product based distance is also used for formulating higher-order losses in Mahalanobis metric learning [41,35]. Given two vectors x and y, a Mahalanobis distance is defined as:…”

Co-Metric Learning for Person Re-Identification

“…Ahmed et al [15] presented a deep convolutional architecture that captured local relationships between person images based on mid-level features. Generally, deep learning is usually utilized to learn feature representations by using deep convolutional features [14][15][16][17] or from the fully connected features [18][19][20] in person re-identification works.…”

“…(i) Feature construction and learning aim at designing or studying discriminative appearance descriptions [8][9][10][11][12][13][14][15][16][17][18][19][20] that are robust for distinguishing different pedestrians across arbitrary cameras. However, handcrafted feature construction is extremely challenging due to miscellaneous and complicated variations.…”