Pedestrian re-recognition method based on domain invariant information separation of low-rank prior guidance

Abstract

The invention relates to a pedestrian re-recognition method based on domain invariant information separation of low-rank prior guidance, and belongs to the field of computer vision. And embedding thedomain invariant information into a dictionary learning framework, and constructing a cross-data-set unsupervised pedestrian re-recognition discrimination dictionary learning model. According to the low-rank priori property of style information, the model can separate domain information aliasing in pedestrian image features from domain invariant information reflecting the pedestrian features; meanwhile, in consideration of domain invariance of pedestrian attributes, the attributes serve as links between domains and are used for constructing the relation between the source data set and the target data set, and domain offset between the source data set and the target data set is reduced. And finally, finely adjusting previously learned parameters through a self-training strategy. Experimentsshow that the method approaches or even exceeds the performance of supervised non-deep learning and unsupervised domain self-adaptive pedestrian re-recognition based on deep learning on many data sets.

Description

technical field [0001] The invention relates to a pedestrian re-identification method based on domain-invariant information separation guided by a low-rank prior, and belongs to the field of computer vision. Background technique [0002] Pedestrian re-identification is a technique to search for the same pedestrian picture from multiple pedestrian pictures under different cameras. Since this technique plays an important role in intelligent surveillance, it has attracted great attention in both academia and industry. In the actual monitoring environment, the images of pedestrians captured by the camera often have low resolution. At the same time, due to the difference in viewing angle and the change of illumination, pedestrians often show strong appearance ambiguity under different viewing angles. great challenge. Although the performance of person re-identification based on deep learning has been significantly improved in recent years, most of these methods are supervised l...

AI Technical Summary

Problems solved by technology

However, the design methods of unsupervised domain-invariant features are often difficult to mine information-rich discriminative information from the data distribution; domain adaptation methods based on adversarial learning often achieve domain alignment through adversarial learning between different domain features, but in In the process of adversarial learning, the extracted features are often shared information from samples in different domains, and the unique information between samples in different domains is lost, which easily aggravates the ambiguity between sample features

Although the method based on image style transfer is effective, it is easy to cause the label drift of the transferred image

In addition to pedestrian re-identification, the method of unsupervised domain adaptation has also received extensive attention and achieved significant research progress, but these methods often assume that the source domain and the target domain have the same or part of the same class. This assumption and The situation of pedestrian re-identification does not match

In addition, in unsupervised domain adaptation methods, the relationship between the source domain and the target domain is often a single-domain vs. single-domain problem

However, in person re-identification, both the labeled source data set and the unlabeled target data set often contain multiple camera views (each camera view can be regarded as a domain), so the method of unsupervised domain adaptation cannot directly Promoted and applied to pedestrian re-identification

Images