Human body key point detection model generation method, system and device and storage medium

Abstract

The invention discloses a human body key point detection model generation method. The method comprises the steps of obtaining a human body posture image; dividing the human body posture images markedwith the key points in advance into a training set and a test set, wherein the key points are body parts of a human body; inputting the training set into a pre-constructed human body key point detection model to train the human body key point detection model, wherein the human body key point detection model is a deep convolutional neural network of RestNet based on embedded seNet, and the human body key point detection model performs Gaussian distribution processing on an input training set to output a loss value of a loss function of the human body key point detection model; and when the lossvalue reaches a preset convergence range, outputting a human body key point detection model. The invention discloses a human body key point detection model generation system and device and a computerreadable storage medium. By adopting the human body key point detection model generated by the embodiment of the invention, the human body key point detection efficiency and accuracy can be improved.

Description

technical field [0001] The invention relates to the technical field of human body detection, in particular to a method, system, device and storage medium for generating a key point detection model of a human body. Background technique [0002] During or after law enforcement, the public security organs need to conduct remote video surveillance on suspects in different scenarios (such as information collection room, interrogation room, interrogation room, detention center, etc.) to ensure that the suspect can record the scene Abnormal behaviors of suspects (such as fighting, falling to the ground, etc.) can be monitored and alarmed in real time. In the prior art, human body key point detection algorithms are used to monitor human body posture. At present, there are two main types of human body key point algorithms: the first type is a top-down (top-down) method, which is used in human body detection The key point detection is carried out one by one on the basis of the human ...

AI Technical Summary

Problems solved by technology

Under the same conditions, the detection effect of the first type of method is obviously better than that of the second type of method, but the general model of the existing type of algorithm is relatively complex, the video memory occupies a large amount, and the detection speed is greatly affected by the number of people in the scene.

Images