Pose recognition method and system based on pose recognition neural network

By constructing a dual-branch fusion network that combines graph convolution and global vision branches, the accuracy and safety assessment issues of existing pose recognition methods in complex scenarios are solved, enabling efficient identification and assessment of dangerous poses in industrial settings.

CN121259912BActive Publication Date: 2026-04-21YUFENG CULTURE TECHNOLOGY (NANTONG) CO LTD
2 Cites 0 Cited by

Patent Information

Application Number
CN202511397024.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-09-28
Publication Date
2026-04-21
Estimated Expiration
2045-09-28

AI Technical Summary

Technical Problem

Existing posture recognition methods are easily affected by occlusion, changes in lighting, and background interference in complex scenes, making it difficult to accurately identify the postures of multiple people in the same frame or with overlapping postures. Furthermore, they lack a comprehensive assessment of the risks associated with postures, failing to meet the needs of safety monitoring and health management in practical applications.

Method used

A dual-branch fusion network based on graph convolution and global vision is constructed. The graph convolution branch models the topology of human key points, and the global vision branch extracts contextual information. The post-processing judgment of geometric rule differences is introduced to identify and evaluate the safety and danger of the pose.

Benefits of technology

It significantly improves the accuracy and robustness of recognizing complex postures, reduces the false alarm rate of hazards, and provides highly intelligent and reliable safety assurance, making it suitable for industrial safety production.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention provides a posture recognition method and system based on a posture recognition neural network, belonging to the field of posture recognition technology. The invention constructs a standard library containing safe and dangerous posture categories and generates a soft-label sample dataset by annotating several key points. Key points are defined as graph nodes, and feature matrices and adjacency matrices are constructed. A dual-branch fusion network model is designed, extracting topological features of key points through a graph convolution branch and extracting global contextual features of the image through a global vision branch. After fusion, a posture similarity spectrum vector is output. Finally, the posture state is determined by calculating the geometric difference between the image to be analyzed and the most similar posture template in terms of key point distance and joint angles, and comparing its overall difference with the safe / dangerous categories. This invention combines the advantages of data-driven and rule-based validation, significantly improving the robustness of dangerous posture recognition and the interpretability of decisions, effectively reducing safety risks.
Need to check novelty before this filing date? Find Prior Art