Gesture recognition method, system and device based on surface electromyography signals

By training a gesture recognition model based on surface electromyography signals using convolutional neural networks and federated averaging algorithms, the problems of cross-domain issues and insufficient data were solved, achieving high-accuracy recognition even under data-scarce conditions, protecting data privacy, and improving the model's adaptability.

CN115840505BActive Publication Date: 2026-07-24UNIV OF SCI & TECH BEIJING +1
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
UNIV OF SCI & TECH BEIJING
Filing Date
2022-12-06
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing technologies for gesture recognition based on surface electromyography signals suffer from problems such as large differences between subjects and between conversations, and insufficient ability to adapt to new domains when data is insufficient, making it difficult to guarantee recognition accuracy.

Method used

The model is trained using a convolutional neural network (CNN) combined with a federated average algorithm (FedAvg), and accuracy is improved across domains through transfer learning. The model is trained using several small databases, combined with data from multiple clients to form a joint model, and the parameters are fine-tuned in the target domain.

Benefits of technology

It effectively reduces cross-domain impact in situations of data scarcity, improves the accuracy and generalization ability of identification, protects data privacy, and enables the network to be trained in a short time and perform well in the target domain.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115840505B_ABST
    Figure CN115840505B_ABST
Patent Text Reader

Abstract

The application provides a gesture recognition method, system and device based on surface electromyogram signals, the method comprising: initializing a model in a server, establishing a joint model; a client collects local data; the server broadcasts the joint model to the client; the client trains the joint model using its local data on the client side, forming a client model; the parameter matrix of the client model is uploaded to the server; the server obtains the parameter matrix of a new joint model based on the parameter matrix; after reaching a preset update round, a final joint model is obtained. This scheme can effectively reduce the influence of cross-domain under the condition of data scarcity, combine multiple clients with small data sets, train a joint model with strong generalization ability under the premise of protecting data privacy, and when new data is encountered, parameter fine-tuning is performed on the joint model, so that a network model with good performance on new data can be obtained in a short time.
Need to check novelty before this filing date? Find Prior Art