3D human body posture estimation model training method

Abstract

The invention discloses a 3D human body posture estimation model training method, which belongs to the technical field of intersection of digital image processing and machine learning, and comprises the following steps of constructing an attitude estimation network model, acquiring a target 3D sample image, setting anchor points on the sample image at a fixed interval, inputting an image matrix into the network model, and acquiring a trained posture estimation model after multiple rounds of iterative training, and using the trained network model to perform posture estimation on the image containing the posture of a human body. According to the 3D human body posture estimation model training method, the coordinates of the articulation points of the human body are predicted by adopting a method of spatially setting the anchor points, so that very high accuracy rate can be achieved; meanwhile, the calculation amount is much lower than that of traditional 3D attitude estimation, and the 3Dhuman body posture estimation model training method has a certain application prospect.

Description

technical field [0001] The invention belongs to the cross technical field of digital image processing and machine learning, and more specifically relates to a 3D human body pose estimation model training method. Background technique [0002] With the development of computer vision technology, the pose estimation technology for depth map and RGB map has developed rapidly. Pose estimation is widely used as a basic technology in human-computer interaction, augmented reality, human behavior analysis, medical rehabilitation, games and other fields. The following are the commonly used pose estimation algorithms: [0003] Existing human pose estimation methods are generally divided into three categories: model fitting methods, discriminative methods, and hybrid methods. The model fitting method uses an optimization method to fit a predefined hand model to the input depth image. The discriminative method is completely data-driven, and its goal is to learn a regressor through labe...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the problem that the existing 3D human body posture estimation method is difficult to obtain high-precision 3D joint point coordinates of the human body, and the time complexity is high, the present invention proposes a method using spatial anchor points with high accuracy and reduced calculation amount. 3D Human Pose Estimation Model Training Method

Images