Pedestrian re-identification method and device, electronic equipment, storage medium, program product

Abstract

The embodiment of the present application discloses a method and device for pedestrian re-identification, electronic equipment, storage media, and program products to obtain images to be identified and candidate image sets; use a feature extraction network to perform feature extraction on each candidate image in the image to be identified and candidate image sets , to obtain the intermediate features to be recognized corresponding to the image to be recognized and the candidate intermediate features corresponding to the candidate image, the feature extraction network is obtained through cross-modal training of image features and language description; The recognition result corresponding to the recognition image uses the natural correspondence between the image and the language describing the image, and further mines the correlation between the local image area and the noun phrase by means of phrase reconstruction, which strengthens the constraints on image feature learning. The quality of visual features for pedestrian re-identification is improved, which in turn improves the accuracy of pedestrian re-identification.

Description

technical field [0001] This application relates to computer vision technology, especially a pedestrian re-identification method and device, electronic equipment, storage media, and program products. Background technique [0002] Pedestrian re-identification is a key technology in intelligent video surveillance systems. It aims to find target samples in a large number of candidate samples by measuring the similarity between a given target sample and candidate samples. Visual features for person re-identification are enhanced with the application of deep neural networks. In order to further improve the discriminative ability of features, some methods use auxiliary data; but there are still the following problems: relying on additional equipment or models, increasing the cost of algorithm running price and time; or defining complex labeling formats for auxiliary data, which increases Human cost of data labeling. Contents of the invention [0003] A pedestrian re-identificat...

AI Technical Summary

Problems solved by technology

In order to further improve the discriminative ability of features, some methods use auxiliary data; but there are still the following problems: relying on additional equipment or models, increasing the cost of algorithm running price and time; or defining complex labeling formats for auxiliary data, which increases Human cost of data labeling

Images