Attitude estimation and network training method, device and system, and storage medium

Abstract

Embodiments of the present invention provide a training method, device and system for a pose estimation network, a pose estimation method, device and system, and a storage medium. The training method includes: obtaining N 1 sample images and labeled data; using N 1 A sample image and labeled data are used to train the pose estimation network, where N 1 The training of the pose estimation network with sample images and labeled data includes: performing the second operation in each iteration of the training process. Using the second operation can focus on optimization for difficult samples. Therefore, the detection accuracy of joint points can be effectively improved.

Description

technical field [0001] The present invention relates to the field of image processing, and more particularly relates to a training method, device, system and storage medium of a pose estimation network, and a pose estimation method, device, system and storage medium. Background technique [0002] Human pose estimation is an important branch of computer vision, which is used to locate various joint points of the human body, such as elbows, wrists, knees, etc. In recent years, with the development of neural networks and hardware facilities, human joint detection technology has made great progress. However, for some difficult samples with large motion range, strange or rare poses, and occlusions, the existing algorithms are generally ineffective, and the detection accuracy needs to be further improved. Contents of the invention [0003] The present invention has been made in consideration of the above-mentioned problems. The present invention provides a posture estimation n...

AI Technical Summary

Problems solved by technology

However, for some difficult samples with large range of motion, strange or rare poses, and occlusions, the existing algorithms are generally ineffective, and the detection accuracy needs to be further improved.

Images