Geethanjali P, Ray K K
School of Electrical Engineering, VIT University, Vellore 632 014, Tamil Nadu, India.
Australas Phys Eng Sci Med. 2011 Sep;34(3):419-27. doi: 10.1007/s13246-011-0079-z. Epub 2011 Jun 11.
The authors in this paper propose an effective and efficient pattern recognition technique from four channel electromyogram (EMG) signals for control of multifunction prosthetic hand. Time domain features such as mean absolute value, number of zero crossings, number of slope sign changes and waveform length are considered for pattern recognition. The patterns are classified using simple logistic regression (SLR) technique and decision tree (DT) using J48 algorithm. In this study six specific hand and wrist motions are identified from the EMG signals obtained from ten different able-bodied. By considering relevant dominant features for pattern recognition, the processing time as well as memory space of the SLR and DT classifiers is found to be less in comparison with neural network (NN), k-nearest neighbour model 1 (kNN-Model-1), k-nearest neighbour model 2 (kNN-Model-2) and linear discriminant analysis. The classification accuracy of SLR classifier is found to be 91 ± 1.9%.
本文作者提出了一种高效有效的模式识别技术，用于从四通道肌电图（EMG）信号控制多功能假手。模式识别考虑了诸如平均绝对值、过零次数、斜率符号变化次数和波形长度等时域特征。使用简单逻辑回归（SLR）技术和采用J48算法的决策树（DT）对模式进行分类。在本研究中，从十个不同的健全人获取的EMG信号中识别出六种特定的手部和腕部动作。通过考虑模式识别的相关主导特征，发现与神经网络（NN）、k近邻模型1（kNN-模型-1）、k近邻模型2（kNN-模型-2）和线性判别分析相比，SLR和DT分类器的处理时间以及内存空间更少。发现SLR分类器的分类准确率为91±1.9%。
