Pedestrian re-identification method based on double constraint metric learning and sample reordering

A technology of pedestrian re-identification and metric learning, applied in the field of pedestrian re-identification based on double-constrained metric learning and sample reordering, which can solve the problems of only considering cross-camera correlation information and ignoring the correlation of different pedestrian pictures.

CN107145826AActive Publication Date: 2017-09-08ZHEJIANG UNIV
7 Cites 18 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2017-09-08

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a pedestrian re-identification method based on double constraint metric learning and sample reordering. The method comprises two stages of training and testing; the training stage comprises the following steps: establishing a cross-camera correlation constraint; establishing a same-camera correlation constraint; and solving a metric matrix; the testing stage comprises the following steps: using the metric matrix to perform feature space projection; calculating the Euclidean distance of query pictures and candidate pictures in a feature space; calculating the initial ordering of the candidate pictures; selecting the first K candidate pictures in a ordering queue; constructing a probabilistic hypergraph by using the relevance of the first K candidate pictures in the feature space; calculating a reordering result based on the probabilistic hypergraph; and returning the final ordering of the candidate pictures. The pedestrian re-identification method based on the double constraint metric learning and sample reordering provided by the invention considers two correlation constraints of training samples simultaneously, so that a feature space obtained by learning is more suitable for pedestrian re-identification, and at the same time, the relevance of the candidate pictures is used to reorder, so that a more accurate pedestrian re-identification result is obtained.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to a method in the technical field of video image processing, in particular to a pedestrian re-identification method based on double-constrained metric learning and sample reordering. Background technique

[0002] Video surveillance provides a rich source of information for security early warning, investigation and evidence collection, and suspect tracking. However, the monitoring range of a single camera is very limited, so it is impossible to carry out all-round monitoring of larger or more complex scenes (such as train stations, airports, campuses, etc.). In order to capture more comprehensive and extensive information in public areas, a large number of surveillance cameras are usually required to work together. The traditional video processing technology is mainly designed for a single camera. When the pedestrian target moves out of the current video, it is impossible to determine the whereabouts of the target. Therefore, ho...

Examples

Embodiment

[0076] In this embodiment, pedestrian images captured by different cameras are processed, a metric matrix is ​​learned through the training set, and a query image of a certain pedestrian target is used in the test phase to find the correct matching of pedestrian targets in the candidate sets captured by different cameras. figure 1 , in an embodiment of the present invention, including two stages of training and testing;

[0077] The training phase includes the following steps:

[0078] Step 1. Establish cross-camera association constraints: Use pedestrian images from different cameras in the training set to form cross-camera sample pairs, and establish constraints so that the feature distance between cross-camera positive sample pairs is smaller than the cross-camera negative sample pair. Feature distance between pairs , which includes the following sub-steps:

[0079] Step 1.1, define training images from different cameras as query sets and candidate set where x i and y...