Recognition of sketching from surface electromyography

Abstract

The main objective of this study is to recognize sketching precisely and effectively in human computer interaction. A surface electromyography (sEMG) based sketching recognition method is proposed. We conducted an experiment in which we recorded the sEMG signals from the forearm muscles of two participants who were instructed to sketch seven basic one-stroke shapes. Subsequently, seven features of the sEMG time domain were extracted. After reducing data dimensionality using principal component analysis, these features were used as input vectors to a sketching recognition model based on Support Vector Machines (SVM). The performance of this model was compared to two other recognition models based on Multilayer perceptron (MLP) neural networks and Self Organization Feature Map (SOFM) neural networks. The average recognition rates for the seven basic one-stroke shapes of two participants achieved by the SVM-based and MLP-based models were both 98.5% in the test set, which were slightly superior to the performance of the SOFM classifier. Our results demonstrate the feasibility of converting forearm sEMG signals into sketching patterns.

Human–Machine Interfaces Based on Bioelectric Signals: A Narrative Review with a Novel System Proposal

“…Huang et al [13] proposed an EMG-based handwriting recognition method based on dynamic time warping and applied it to the recognition of digits, Chinese numerals, and English letters from the six-channel EMG signals. Chen et al [14] proposed a sketching recognition method using surface EMG from forearm muscles.…”

Myoelectric Interfaces and Related Applications: Current State of EMG Signal Processing–A Systematic Review

“…• Surface EMG-based sketching recognition for the development of electronic sketching systems and computeraided sketching systems [46], [47].…”

“…• 500 Hz, 3rd order [22] • 5-500 Hz, 2nd order [41] • 10-200 Hz [44] • 10-400 Hz [4], 3rd order [33], [4] • 10-500 Hz [30], [46], [47] • 20-250 Hz 2nd order [42] • 20-450 Hz [49] • 20-500 Hz [48], 2nd order [23] and 4th order [31], [32] • 30-300 Hz 4th order [29] • 2-750 Hz [37] • 13 works do not specify the characteristics of the filtering stage (2) Notch filters: 50 Hz and 60 Hz (depending on the frequency of the electrical network in each country).…”

“…Regarding the sampling frequency, 24 out of the 30 works from Table 5 presented a detailed report of the sampling frequency considered in their applications, summarized as follows: 21 papers reported sampling frequency values ranging from 1000 to 1500 samples per second; for the remaining 3 papers, the sampling frequency ranges from 2000 to 23434 samples per second. The latter are related to applications that resort to classification methods for pathology [28]; 12 bits Pathology detection muscular activation [23], [21], Exoskeletons [26], Prosthetic control [30]; 14 and 16 bits Handwriting recognition [46], Prosthesis control [32] and 24 bits Prosthetic control [29], [33].…”