Single-target human body key point detection method, system and device and medium

Abstract

The invention provides a human body key point detection method, system and device based on a single target, and a medium. The method comprises the steps of obtaining one or more initial single targetimages; performing scene enhancement on the obtained one or more initial single-target images, performing learning training on a single-target human body key point detection algorithm based on the single-target images after scene enhancement, and forming a single-target human body detection model; and acquiring human body key points in one or more initial single-target images and corresponding visibility information according to the generated single-target human body detection model. According to the invention, human body key point detection can be carried out on the single-target image, optimization can be carried out in complex scenes such as blurring, shielding, truncation, abnormal postures and the like, and the precision of human body key point detection in the complex scenes is improved. Moreover, models and operators with small parameter quantities and calculated quantities can be adopted, model reasoning is accelerated by utilizing model quantification, and the reasoning time consumption is reduced.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method, system, device and medium for detecting key points of a single-target human body. Background technique [0002] In application scenarios such as pedestrian video structuring and pedestrian re-identification, after obtaining human body detection results, it is necessary to further obtain human body posture information for subsequent calculation of human behavior, human integrity and other related information. Human body pose estimation needs to determine the positions of multiple joints such as the human head, torso, hands, and feet. Therefore, the present invention proposes a way to describe the posture of the human body in the form of key points of the human body, wherein a key point of the human body represents a joint of the human body. Since the position of a single human body area can be obtained through the human body detection module, it is necessary t...

AI Technical Summary

Problems solved by technology

However, the existing mainstream human body key point algorithm model is large, and the post-processing takes a long time, which is not suitable for the actual video surveillance scene that requires fast response; and the blurring, occlusion, truncation, abnormal posture and other situations in the actual scene will It brings difficulty to the detection of key points of the human body

Images