Christine Serviere

EMG feature evaluation for improving myoelectric pattern recognition robustness

In pattern recognition-based myoelectric control, high accuracy for multiple discriminated motions is presented in most of related literature. However, there is a gap between the classification accuracy and the usability of practical applications of myoelectric control, especially the effect of long-term usage. This paper proposes and investigates the behavior of fifty time-domain and frequency-domain features to classify ten upper limb motions using electromyographic data recorded during 21days. The most stable single feature and multiple feature sets are presented with the optimum configuration of myoelectric control, i.e. data segmentation and classifier. The result shows that sample entropy (SampEn) outperforms other features when compared using linear discriminant analysis (LDA), a robust classifier. The averaged test classification accuracy is 93.37%, when trained in only initial first day. It brings only 2.45% decrease compared with retraining schemes. Increasing number of features to four, which consists of SampEn, the fourth order cepstrum coefficients, root mean square and waveform length, increase the classification accuracy to 98.87%. The proposed techniques achieve to maintain the high accuracy without the retraining scheme. Additionally, this continuous classification allows the real-time operation.

Feature extraction of the first difference of EMG time series for EMG pattern recognition

This paper demonstrates the utility of a differencing technique to transform surface EMG signals measured during both static and dynamic contractions such that they become more stationary. The technique was evaluated by three stationarity tests consisting of the variation of two statistical properties, i.e., mean and standard deviation, and the reverse arrangements test. As a result of the proposed technique, the first difference of EMG time series became more stationary compared to the original measured signal. Based on this finding, the performance of time-domain features extracted from raw and transformed EMG was investigated via an EMG classification problem (i.e., eight dynamic motions and four EMG channels) on data from 18 subjects. The results show that the classification accuracies of all features extracted from the transformed signals were higher than features extracted from the original signals for six different classifiers including quadratic discriminant analysis. On average, the proposed differencing technique improved classification accuracies by 2-8%.

A novel method for permutation correction in frequency-domain in blind separation of speech mixtures

This paper presents a method for blind separation of convolutive mixtures of speech signals, based on the joint diagonalization of the time varying spectral matrices of the observation records and a novel technique to handle the problem of permutation ambiguity in the frequency domain. Simulations show that our method works well even for rather realistic mixtures in which the mixing filter has a quite long impulse response and strong echoes.

A feasibility study on the use of anthropometric variables to make muscle–computer interface more practical

High classification accuracy has been achieved for muscle–computer interfaces (MCIs) based on surface electromyography (EMG) recognition in many recent works with an increasing number of discriminated movements. However, there are many limitations to use these interfaces in the real-world contexts. One of the major problems is compatibility. Designing and training the classification EMG system for a particular individual user is needed in order to reach high accuracy. If the system can calibrate itself automatically/semi-automatically, the development of standard interfaces that are compatible with almost any user could be possible. Twelve anthropometric variables, a measurement of body dimensions, have been proposed and used to calibrate the system in two different ways: a weighting factor for a classifier and a normalizing value for EMG features. The experimental results showed that a number of relationships between anthropometric variables and EMG time-domain features from upper-limb muscles and movements are statistically strong (average r=0.71−0.80) and significant (p<0.05). In this paper, the feasibility to use anthropometric variables to calibrate the EMG classification system is shown obviously and the proposed calibration technique is suggested to further improve the robustness and practical use of MCIs based on EMG pattern recognition.

A phase-based dual microphone method to count and locate audio sources in reverberant rooms

This paper presents an off-line method to estimate the mixing conditions, characterized by the number of audio sources and their time difference of arrival (TDOA). The proposed method is based on the assumption of having, at least, one time frame where each one of the sources is dominant. From such frames, the TDOA of the correspondent dominant source is estimated by maximizing a novel coherence function similar to the linear regression proposed in. Finally, the number of sources will be derived by counting the number of active directions of arrival (DOA). Experimental results on real-life recordings are presented, showing the good performance of the algorithm in reverberant environments.

The biomechanical model of the long finger extensor mechanism and its parametric identification

The extensor mechanism of the finger is a structure transmitting the forces from several muscles to the finger joints. Force transmission in the extensor mechanism is usually modeled by equations with constant coefficients which are determined experimentally only for finger extension posture. However, the coefficient values change with finger flexion because of the extensor mechanism deformation. This induces inaccurate results for any other finger postures. We proposed a biomechanical model of the extensor mechanism represented as elastic strings. The model includes the main tendons and ligaments. The parametric identification of the model in extension posture was performed to match the distribution of the forces among the tendons to experimental data. The parametrized model was used to simulate three degrees of flexion. Furthermore, the ability of the model to reproduce how the force distribution in simulated extensor mechanism changes according to the muscle forces was also demonstrated. The proposed model could be used to simulate the extensor mechanism for any physiological finger posture for which the coefficients involved in the equations are unknown.

Modeling the Short Time Fourier Transform Ratio and Application to Underdetermined Audio Source Separation

A process for making a known 6-methyl, 19-nor-pregna-4, 6-diene, 3.20-dione which begins with formylating a 3-alcoxy, 19-nor-pregna-3,5,17(20)-triene at the 6 position. The 6-formylated derivative is reduced to yield a 6-hydroxy methylated derivative, which is in turn dehydrated to a 3-keto, 6-methylenic derivative. The 3-keto derivative is then isomerized to a 3-keto, 4,6,17-pregnatriene. This latter triene is then coverted to the known product by reaction with a bis-hydroxylating agent and a catalyst based on osmium tetroxide. Optionally, the product can be acylated at the 17-alpha position. The process reduces the cost of producing the known product by allowing it to be manufactured from starting materials less costly than those previously required.

Extraction of EI and EDM muscle sources from surface electromyographic signals using delay estimation

Interference or a crosstalk from nearby muscles is a classic problem in surface electromyographic recordings. Most studies of crosstalk diminution have only focused on blind source separation techniques and spatial filtering of simulated signals. The aim of this study was to perform an extraction of the activity of index and little finger extensor muscles from the electromyographic mixtures recorded with a surface electrode array. The motor unit action potential were detected in the mixtures and classified according to an inter-electrode delay. The signal-to-interference ratio performance of the proposed method was higher than performance of the beamformer and the frequency JADE algorithms. The findings show that creation of a muscle action potential propagation model, required by beamformer, could be problematic. The proposed algorithm uses a generalized model and shows higher performance.

Blind Separation of Surface Electromyographic Mixtures from Two Finger Extensor Muscles

Blind source separation BSS was performed to reduce the crosstalk in the surface electromyografic signals SEMG for the muscle force estimation applications. A convolutive mixture model was employed to separate the SEMG signals from two finger extensor muscles using a frequency-domain approach. It was assumed that the tension of each muscle varies independently and the independence of the SEMG was replaced by minimization of the covariance of muscle forces represented by integrated SEMG. This covariance was also used to resolve the permutation ambiguity inherent to the frequency-domain BSS. The forces estimated by the reconstructed sources were compared with the measured forces to calculate the crosstalk reduction efficiency. The proposed algorithm was shown to be more effective in frequency domain than an ICA algorithm for extensor muscles crosstalk reduction.

Generalization of GLRT-based magnetic anomaly detection

Magnetic anomaly detection (MAD) refers to a passive method used to reveal hidden magnetic masses and is most commonly based on a dipolar target model. This paper proposes a generalization of the MAD through a multipolar model that provides a more precise description of the anomaly and serves a twofold objective: to improve the detection performance, and to widen the variety of detectable targets. The dipole detection strategy - namely an orthonormal decomposition of the anomaly followed by a generalized likelihood ratio test - is hence revisited in the multipolar case. The performance are assessed analytically and the relevance of this generalization is demonstrated on multipolar scenarios.