Unsupervised human body key point detection method and system in fixed scene

Abstract

The invention relates to an unsupervised human body key point detection method and system in a fixed scene, and the method comprises the steps of training a convolutional neural network through employing a plurality of character color image samples synchronously shot by color cameras at different shooting angles in a camera array; inputting the color images of the to-be-detected person at different shooting angles of view into an encoder of the trained convolutional neural network to obtain heat maps of the to-be-detected person at different shooting angles of view; and extracting three-dimensional coordinates of human body key points of the to-be-detected person from the heat maps of the to-be-detected person at different shooting angles of view. According to the invention, unsupervised training is carried out on the convolutional neural network, the problem that training of the neural network needs a large amount of labeled data is solved, accurate measurement of the three-dimensional coordinates of the human body key points is realized by using the convolutional neural network, and the problem that a marker needs to be worn during traditional human body key point monitoring is solved.

Description

technical field [0001] The invention relates to the technical field of human body key point detection, in particular to an unsupervised human body key point detection method and system in a fixed scene. Background technique [0002] Human body key point detection plays an important role in human-computer interaction, virtual reality, and movie special effects production. At present, most of the human body key point detection schemes adopt one of the following two ideas: 1. Put on some distinctive markers by the characters, and then use sensors or feature detection algorithms to detect the spatial coordinates of the markers; 2. Abandon the traditional Based on the sensor and feature detection algorithm, the key points of the human body in a large number of color images are manually marked in advance, and then the neural network is trained with a large amount of data, and finally the key points of the person are detected using the trained neural network. However, both methods...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide an unsupervised human body key point detection method and system in a fixed scene, using unsupervised neural network technology to realize accurate measurement of three-dimensional coordinates of human body key points, to solve the need to wear markers when traditional human body key point monitoring The problem and the problem that the training of the neural network requires a large amount of labeled data

Images