EMG-Based Cross-Subject Silent Speech Recognition Using Conditional Domain Adversarial Network
- 18 September 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 15 (4) , 2282-2290
Abstract
Machine learning techniques have achieved great success in electromyography (EMG) decoding, but EMG-based cross-subject silent speech recognition (SSR) received less attention because of its high individual variability. Therefore, this paper explores the field of cross-subject SSR to improve the recognition performance of EMG data collected from new subjects. Firstly, this paper reports on applying time-series features and one-dimensional convolutional neural networks (1D-CNN) for cross-subject SSR. Secondly, this paper proposes using a conditional domain adversarial network (CDAN) to solve the problem of reduced cross-subject SSR accuracy in the few samples’ datasets. It innovatively integrates the maximum mean difference (MMD) loss to get an improved condition domain adversarial network (ICDAN). While 1D-CNN is a feature extraction network that can meet the needs of cross subject silent speech recognition in large datasets, the recognition effect will be weakened in small datasets. Adding an ICDAN network after the feature extraction network can improve the problem of data distribution differences between the two domains, and further enhance recognition performance. The results show that the 1D-CNN model based on time-series features yields better results in the SSR of new subjects, and the ICDAN model can further improve the classification accuracy of cross-subjects in a few sample datasets by 14.88%.Keywords
Funding Information
- National Natural Science Foundation of China (61703407, 61901505, 62076250)
This publication has 30 references indexed in Scilit:
- Silent Speech Recognition as an Alternative Communication Device for Persons With LaryngectomyIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2017
- Preface: Recent Advances in Understanding the Human Phonatory ProcessActa Acustica united with Acustica, 2016
- Surface electromyographic activity of non-laryngeal neck muscles in Cantonese tone production2014
- Fluid Dynamics of Human Phonation and SpeechAnnual Review of Fluid Mechanics, 2013
- Signal acquisition and processing techniques for sEMG based silent speech recognition2011
- Silent speech interfacesSpeech Communication, 2010
- Modeling coarticulation in EMG-based continuous speech recognitionSpeech Communication, 2010
- Session independent non-audible speech recognition using surface electromyography2005
- Hidden Markov model classification of myoelectric signals in speechIEEE Engineering in Medicine and Biology Magazine, 2002
- Research summary of a scheme to ascertain the availability of speech information in the myoelectric signals of neck and head muscles using surface electrodesComputers in Biology and Medicine, 1986