Pedestrian attribute identification method and device, storage medium and electronic device

Abstract

The embodiment of the invention provides a pedestrian attribute identification method and device, a storage medium and an electronic device. The pedestrian attribute recognition method comprises the steps that a pedestrian image to be recognized is acquired, pedestrian local attributes and pedestrian global attributes of the pedestrian image to be recognized are recognized through a pre-trained pedestrian attribute recognition model, and in the pre-training process, the pedestrian local attributes and the pedestrian global attributes of the pedestrian image to be recognized are recognized by calculating the loss between an edge pedestrian local subgraph and a basic pedestrian local subgraph. And pre-training the pedestrian attribute recognition model, and outputting the pedestrian local attribute and the pedestrian global attribute. In the scheme of the embodiment of the invention, as the pre-trained pedestrian attribute recognition model learns the ability to divide the local area of the target object and the ability to detect the target area in the local area of the target object, the recognition accuracy of the pedestrian attributes is improved.

Description

technical field [0001] Embodiments of the present invention relate to the field of computer technologies, and in particular, to a pedestrian attribute identification method, device, storage medium, and electronic device. Background technique [0002] With the widespread popularity of video surveillance and the continuous improvement of deep learning technology, the identification and classification of pedestrian attributes are widely used in the field of security. [0003] Long-term historical experience in the security field has shown that the human body image data collected by the camera is an effective detective clue. Extracting biometric features based on video data and carrying out personal identification can greatly improve the accuracy and effectiveness of public security disposal. Traditional video analysis is to carry out identification by extracting facial features. However, it is difficult to capture frontal, unobstructed and clear faces in real-life snapshot data...

AI Technical Summary

Problems solved by technology

This process is not only not concise enough, but also causes a waste of computing and storage resources in the judgment process

[0005] Moreover, there are many attributes of pedestrians, and it is often impossible to learn the attribute characteristics distributed in various positions of pedestrians only through simple network classification. Through segmentation, the method of dividing the analogy required by each position will cause a large number of manual labeling tasks and consume a lot of time. Manpower and time will also cause a waste of computing and storage resources

Images