Pedestrian re-identification method fusing random batch masks and multi-scale representation learning

Abstract

The invention relates to a pedestrian re-identification method fusing random batch masks and multi-scale representation learning. The pedestrian re-identification method comprises the steps of constructing a pedestrian re-identification training network; performing network hyper-parameter adjustment according to preset training parameters to obtain a learning network; shielding multi-scale representation learning and random batch mask branches to obtain a test network, and inputting the test set into the test network to obtain a corresponding test identification result; judging whether the accuracy of the test recognition result is greater than or equal to a preset value or not, if so, inputting the actual data set into the learning network, and otherwise, retraining the network; and finally, shielding multi-scale representation learning and random batch mask branches to obtain an application network, and inputting the query image into the application network to obtain a correspondingidentification result. Compared with the prior art, the method has the advantages that a random batch mask strategy, multi-scale representation learning and loss function joint training are used, moredetailed discrimination features of pedestrian images can be captured, and local important suppressed features are extracted.

Description

technical field [0001] The invention relates to the technical field of image processing for computer pattern recognition, in particular to a pedestrian re-identification method that combines random batch masks and multi-scale representation learning. Background technique [0002] Person Re-identification (PReID) is a technology that uses computer vision technology to determine whether a specific pedestrian exists in an image or video sequence. It is widely considered as a sub-problem of image retrieval. Automatically retrieve the image of the pedestrian across devices. At present, a large number of cameras used in the field of public security in cities have been deployed, almost reaching the level of coverage of tens of meters to hundreds of meters. However, there are still areas that cannot be covered by different cameras. The goal of pedestrian re-identification is to determine where the target found under a camera has gone after leaving the camera's field of view. This i...

AI Technical Summary

Problems solved by technology

[0004] When extracting image features, traditional methods use deep learning models, automatically learn features based on convolutional neural networks, and use attention mechanisms to extract features, but this The method usually only focuses on extracting facial features or other prominent features in the image, and does not extract locally suppressed features such as hands or footsteps, resulting in the inability to effectively extract these locally suppressed important detailed features, that is, The accuracy of subsequent recognition cannot be guaranteed

Images