Two-dimensional human arm pose estimation method based on feature fusion and attention mechanism

By combining a parallel model of CNN and Transformer with feature fusion and multidimensional attention mechanisms, the problem of insufficient information capture in 2D human arm pose estimation is solved, improving the accuracy and robustness of joint detection.

CN116758636BActive Publication Date: 2026-04-28ZHEJIANG UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202310732180.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-06-20
Publication Date
2026-04-28
Estimated Expiration
2043-06-20

AI Technical Summary

Technical Problem

Existing two-dimensional human arm pose estimation models are difficult to effectively capture global information, and CNNs have limited receptive fields, making it difficult to fully utilize low-level detailed features and high-level semantic features.

Method used

A two-dimensional human arm pose estimation method based on feature fusion and attention mechanism is adopted. The joint feature is extracted by combining a parallel model of CNN and Transformer. The detection capability of joint position is enhanced by feature fusion and multi-dimensional attention mechanism.

Benefits of technology

It improves the ability to detect occluded joints, reduces feature loss, enhances the utilization of high-level semantic information, and improves the model's accuracy in detecting joint positions.

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

Abstract

The present application relates to the field of computer vision, especially to a two-dimensional human arm posture estimation method based on feature fusion and attention mechanism, comprising the following steps: step one, collecting human action images, and performing image cropping through human boundary range; step two, performing human arm joint feature extraction on the cropped images through a basic network module; step three, performing feature fusion and focusing through a feature fusion and attention combination module, so as to reduce feature loss and improve effective information concentration, and obtaining an arm joint heat map; step four, obtaining position coordinates from the obtained arm joint heat map. The present application can obtain accurate two-dimensional arm joint position coordinates from human action RGB images.
Need to check novelty before this filing date? Find Prior Art