{"title":"Surface Electromyography based Hand Gesture Signal Classification using 1D CNN","authors":"K. S, J. Sahoo, S. Ari","doi":"10.1109/ISACC56298.2023.10083548","DOIUrl":null,"url":null,"abstract":"Surface electromyography-based hand gesture recognition for human-machine interfacing has captivated the interest of researchers in recent years. Machine learning techniques are highly dependent on the features selected for classification. A poor choice of features can deteriorate the classification ability of the system. In order to overcome this issue, deep learning techniques which are capable of automatic feature extraction are employed these days. For any ideal classifying machine, it is required to have a short training period and an ability to produce classification results within a short interval of applying an input. This work proposes a 1D CNN model which is able to classify all the 52 gestures of the NinaPro DB1 database which can be applied in real time applications with fast response. The developed model achieves an accuracy of 78.95% with just 351,532 trainable parameters and an average inference time of 0.258 ms.","PeriodicalId":211409,"journal":{"name":"2023 International Conference on Intelligent Systems, Advanced Computing and Communication (ISACC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2023 International Conference on Intelligent Systems, Advanced Computing and Communication (ISACC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISACC56298.2023.10083548","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Surface electromyography-based hand gesture recognition for human-machine interfacing has captivated the interest of researchers in recent years. Machine learning techniques are highly dependent on the features selected for classification. A poor choice of features can deteriorate the classification ability of the system. In order to overcome this issue, deep learning techniques which are capable of automatic feature extraction are employed these days. For any ideal classifying machine, it is required to have a short training period and an ability to produce classification results within a short interval of applying an input. This work proposes a 1D CNN model which is able to classify all the 52 gestures of the NinaPro DB1 database which can be applied in real time applications with fast response. The developed model achieves an accuracy of 78.95% with just 351,532 trainable parameters and an average inference time of 0.258 ms.