Co-Metric Learning for Person Re-Identification

Abstract

Person re-identification, aiming to identify the same pedestrian images across disjoint camera views, is a key technique of intelligent video surveillance. Although existing methods have developed both theories and experimental results, most of effective ones pertain to fully supervised training styles, which suffer the small sample size (SSS) problem a lot, especially in label-insufficient practical applications. To bridge SSS problem and learning model with small labels, a novel semisupervised co-metric learning framework is proposed to learn a discriminative Mahalanobis-like distance matrix for label-insufficient person re-identification. Different from typical co-training task that contains multiview data originally, single-view person images are firstly decomposed into pseudo two views, and then metric learning models are produced and jointly updated based on both pseudo-labels and references iteratively. Experiments carried out on three representative person re-identification datasets show that the proposed method performs better than state of the art and possesses low label sensitivity.

Re-Identification in Video Surveillance Systems Considering Appearance Changes

“…Also, Global-Local Branches are used to extract the local and global features at the highest level. Also, Leng [28] proposed a semi-supervised cometric learning approach, where a few annotated training samples are used to train a discriminative Mahalanobis-like distance matrix. This approach, rst, uses both the hand-crafted features, i.e., color name [17], and the features extracted from the Siamese CNN [29] for representing single-view person images.…”

“…Training these kinds of networks brings computational and time complexity. The re-identi cation approaches proposed in [14,28,31] used some appearance features such as color naming and features extracted from a deep network. The re-identi cation approach introduced in [14] is better than our IGOG and IHGD in ranks 5, 10, and 20 on the GRID dataset, whereas the IGOG and IHGD outperform in rank 1 on this dataset.…”

COVID-19 contact tracking by group activity trajectory recovery over camera networks

“…However, in real-world surveillance scenario, pedestrians don’t always appear in the camera. Most of the existing methods [28] , [29] , [30] employ CNN to learn discriminative features or design various metric distances for better measuring the similarities between person image pairs. A network consistent re-identification method [31] , [32] considers explicitly maintaining the consistency of the results across the network to obtain the maximal correct matches for the whole camera network.…”