Cross-domain pedestrian re-identification method based on unsupervised joint multi-loss model

Abstract

The invention discloses a cross-domain pedestrian re-identification method based on an unsupervised joint multi-loss model. The method mainly solves the problem of low identification rate of an existing unsupervised cross-domain pedestrian re-identification method. The method comprises the following steps: 1) obtaining a data set, and dividing the data set into a training set and a test set; 2) carrying out a plurality of preprocessings and extensions on the training set; 3) selecting a residual network as a reference network model, initializing network parameters, and adjusting a network structure; 4) constructing a target domain loss function; 5) fusing the target domain loss function with the source domain loss function and the triple loss function to obtain a total loss function; 6) training the residual network by using the total loss function to obtain a trained network model; and 7) inputting the test set into the trained network model, and outputting an identification result. The identification rate of unsupervised cross-domain pedestrian re-identification is improved, the occurrence of an over-fitting situation is effectively avoided, and the method can be used for intelligent security of suspected target search and pedestrian cross-camera tracking.

Description

technical field [0001] The invention belongs to the field of computer vision and deep learning, and specifically relates to a cross-domain pedestrian re-identification method, which can be used in the intelligent security field of searching for suspicious targets and tracking pedestrians across cameras. Background technique [0002] With the construction of digital cities, surveillance videos are more and more widely used in various places where people live. These massive surveillance videos play a very important role in public security fields such as urban security and crime tracking. How to better process these data is also an important issue. Great challenges for our future. [0003] Person re-identification, also known as person re-identification, is a technology that uses computer vision technology to determine whether a specific pedestrian exists in an image or video sequence, and is widely considered as a sub-problem of image retrieval. Given a monitored pedestrian i...

AI Technical Summary

Problems solved by technology

Although this method has achieved good recognition results, due to ignoring the distribution of data features in the target domain itself, the pedestrian re-identification error rate is high under the cross-camera

Images