Pedestrian re-identification method based on key point feature alignment in community monitoring scene

Abstract

The invention relates to the technical field of face detection and deep learning, and particularly discloses a pedestrian re-identification method based on key point feature alignment in a community monitoring scene, which performs multi-angle analysis on pedestrian images or videos in the community scene and performs key point mark alignment on three-dimensional features of a human body trunk inreal life, wherein the BFASM carries out refined feature selection and alignment on pedestrian images under multiple postures, a refined distributed SILTP method establishes feature vectors for refined local features, feature representation of global feature vectors is integrated through weighted calculation, metric learning is carried out on an established global feature vector set, and the Mahalanobis distance judgment function enabling the recognition difference degree to be obvious is automatically obtained. According to the invention, the fine-grained optimization of the local recognitioneffect of the human body is achieved, the pedestrians in the complex community scene can be efficiently recognized on the basis of the conventional pedestrian re-recognition method, and the specificcrowd can be recognized.

Description

technical field [0001] The invention relates to the technical fields of image retrieval, artificial intelligence, and deep learning, and in particular to a pedestrian re-identification method based on key point feature alignment in a community monitoring scene. Background technique [0002] Pedestrian re-identification technology is an important research field in image retrieval, and it is suitable for analyzing and identifying key points of pedestrians in community monitoring scenarios. The analysis results of individual features gradually rise to the overall human body feature matching results to complete Accurate identification of special populations. Pedestrian re-identification technology is one of the optimization directions of pedestrian recognition. It breaks through the limitations of camera resolution, camera angle and pedestrian status, and can obtain effective recognition results for pedestrians in complex environments. [0003] Pedestrian re-identification is a...

AI Technical Summary

Problems solved by technology

Among them, horizontal cutting and rectangular field cutting divide the pedestrian image into different parts for feature matching. However, the "parts" generated by this cutting method cannot accurately reflect the realistic characteristics of the human body.

Segmentation based on parts (hands, feet, head, etc.) can reduce the position error, but the error generated during part matching depends on the design of the specific algorithm

Images