Face attribute association method based on pedestrian re-identification

Abstract

The invention discloses a face attribute association method based on pedestrian re-identification in order to solve the problem of masked disguise identity identification. The method comprises the following steps: firstly, extracting a global feature which is learned together with a local feature, calculating a shortest path between two groups of local features to execute matching, and searching aplurality of pedestrian images which are most similar to a target person; and then, positioning a face position in the corresponding pedestrian image, and quantifying the quality of the face image into a face similarity probability by adopting a fragment-based face image quality evaluation method, thereby selecting a face image with the best identification degree; and finally, establishing a credible identity recognition measurement model, and taking the face image with the highest confidence as an alternative face of the masked target. According to the method, the masked face recognition problem is converted into the correlation corresponding problem between the masked face and the displayed face of the same object, and an effective means is provided for physical identity recognition ofthe masked disguise target.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and relates to a face attribute association method in video data, in particular to a face attribute association method based on pedestrian re-identification. [0002] technical background [0003] Face is the most important basis for identification. Google's FaceNet, the most advanced face recognition technology, has an accuracy rate of 99.63% on the LFW standard data set. However, for the motive of deliberately hiding their identities, criminals often wear sunglasses, masks, hats, turbans and other accessories to cover their faces at the crime scene, or use hoods or even 3D masks to disguise their faces. Behavior poses a huge challenge for identification. [0004] Compared with the influence of factors such as expressions and postures, it is more difficult to deal with facial occlusion factors such as headgear, masks, sunglasses, masks, and hats. In 2016, scholars from Carnegie Mellon U...

AI Technical Summary

Problems solved by technology

[0005] Although the current occluded face recognition has made positive progress, it is limited to a single mild occlusion situation such as glasses, masks, and fake beards. For heavy occlusions such as wearing sunglasses and masks at the same time, the recognition rate will drop sharply. Facial recognition is completely ineffective

Images