Deep Learning for Electromyographic Hand Gesture Signal Classification Using Transfer Learning

Abstract

In recent years, deep learning algorithms have become increasingly more prominent for their unparalleled ability to automatically learn discriminant features from large amounts of data. However, within the field of electromyographybased gesture recognition, deep learning algorithms are seldom employed as they require an unreasonable amount of effort from a single person, to generate tens of thousands of examples.This work's hypothesis is that general, informative features can be learned from the large amounts of data generated by aggregating the signals of multiple users, thus reducing the recording burden while enhancing gesture recognition. Consequently, this paper proposes applying transfer learning on aggregated data from multiple users, while leveraging the capacity of deep learning algorithms to learn discriminant features from large datasets. Two datasets comprised of 19 and 17 able-bodied participants respectively (the first one is employed for pre-training) were recorded for this work, using the Myo Armband. A third Myo Armband dataset was taken from the NinaPro database and is comprised of 10 able-bodied participants. Three different deep learning networks employing three different modalities as input (raw EMG, Spectrograms and Continuous Wavelet Transform (CWT)) are tested on the second and third dataset. The proposed transfer learning scheme is shown to systematically and significantly enhance the performance for all three networks on the two datasets, achieving an offline accuracy of 98.31% for 7 gestures over 17 participants for the CWT-based ConvNet and 68.98% for 18 gestures over 10 participants for the raw EMG-based ConvNet. Finally, a use-case study employing eight able-bodied participants suggests that real-time feedback allows users to adapt their muscle activation strategy which reduces the degradation in accuracy normally experienced over time.

Real-time Continuous Hand Motion Myoelectric Decoding by Automated Data Labeling

“…There are several additional avenues of future work. On the one hand of HMI, HDL shortens the training time with the help of layer-by-layer training and expert experience, but further research on the generalization performance is necessary to make the algorithm adapt to different individuals quickly [26,27] . On the other hand of sEMG controlled prosthesis, we conducted the preliminary test on upper limb amputee and found that the recognition sharply deteriorated before and after wearing the prosthetic hand, which is a difficult problem in the field of prosthetic research [28,29] .…”

Interpreting Deep Learning Features for Myoelectric Control: A Comparison With Handcrafted Features

“…Recent works on sEMG-based gesture recognition using deep learning have shown that ConvNets trained with the raw sEMG signal as input were able to achieve similar classification accuracy to the current state of the art (Zia ur Rehman et al, 2018;Côté-Allard et al, 2019a). Consequently, and to reduce bias, the preprocessed raw data (see Section 2.1) is passed directly as an image of shape 10 × 151 (Channel × Sample) to the ConvNet.…”

“…Using this validation set, learning rate annealing is applied with a factor of five and a patience of fifteen with early stopping being applied when two consecutive annealing occurred without achieving a better validation loss. Dropout is set to 0.35 (following (Côté- Allard et al, 2019a)). Note that all architecture choices and hyperparameters selection were performed using the training set of the 3DC Dataset.…”

Deep Learning in Physiological Signal Data: A Survey

“…Due to lack of datasets, contributions such as Nodera et al [23] employed a technique of data augmentation, in which a fake dataset is generated by duplicating original data and doing a transformation such as translation and rotation. Contributions [12,16,23,46,58] employed a transfer learning technique. Rather than undertaking a training from a scratch with a huge required dataset, they adapted the pre-weight from a state-of-the-art model such as AlexNet, VGG, ResNet, Inception, or DenseNet.…”

“…Hand motion recognition [9][10][11][12][13][14][15][16][17], Muscle activity recognition [18][19][20][21][22][23] ECG Heartbeat signal classification , Heart disease classification [49][50][51][52][53][54][55][56][57][58][59][60][61][62][63], Sleep-stage classification [64][65][66][67][68], Emotion classification [69], age and gender prediction [70] EEG Brain functionality classification , Brain disease classification , Emotion classification [122][123][124][125][126][127][128][129], Sleep-stage classification [130][131][132][133]…”