Human body action recognition method and device based on neural network

Abstract

The embodiment of the invention provides a human body action recognition method and device based on a neural network, and the method comprises the steps: carrying out the preprocessing of an RGB-D image of a to-be-recognized human body action, and obtaining an RGB image without background pixel interference and a point cloud three-dimensional coordinate graph without background pixel interference;inputting the RGB image without background pixel interference and the point cloud three-dimensional coordinate graph without background pixel interference into an attitude parameter identification model, and outputting attitude parameters, morphological parameters and displacement parameters of a to-be-identified human body action, wherein the posture parameter identification model is obtained bytraining a large number of sample labels, inputting posture parameters, morphological parameters and displacement parameters of the human body action to be identified into the parameterized model, and outputting a human body action result to be identified. According to the method and the device provided by the embodiment of the invention, an application scene supporting real-time action migrationis realized, and the accuracy of human body action recognition is improved.

Description

technical field [0001] The invention relates to the technical field of human motion recognition, in particular to a neural network-based human motion recognition method and device. Background technique [0002] The reconstruction and attribute recognition of 3D human body has always been an important research direction in the field of machine vision. At present, the work related to human body reconstruction based on deep learning in academia can be roughly divided into two categories, parametric model reconstruction and non-parametric model reconstruction. Among them, the representative work of non-parametric model reconstruction is Yuncong's DenseBody. This method expands the human body mesh into UVMap, and then performs regression on UVMap through the convolutional network. It has the advantages that data expression is more suitable for convolution and the effect is better. . The representative work of parametric model reconstruction is Berkeley's HMR, which directly uses...

AI Technical Summary

Problems solved by technology

[0006] Embodiments of the present invention provide a neural network-based human motion recognition method and device, which are used to solve the problem that it is difficult to support real-time motion migration application scenarios and the motion capture accuracy is not high in the existing RGB image-based real-time capture human motion method. question

Images