Recently, cardiovascular disease (CVD) has become one of the leading death causes worldwide, and it contributes to 41% of all deaths each year in China. This disease incurs a cost of more than 400 billion US dollars in China on the healthcare expenditures and lost productivity during the past ten years. It has been shown that the CVD can be effectively prevented by an interdisciplinary approach that leverages the technology development in both IT and electrocardiogram (ECG) fields. In this paper, we present WE-CARE , an intelligent telecardiology system using mobile 7-lead ECG devices. Because of its improved mobility result from wearable and mobile ECG devices, the WE-CARE system has a wider variety of applications than existing resting ECG systems that reside in hospitals. Meanwhile, it meets the requirement of dynamic ECG systems for mobile users in terms of the detection accuracy and latency. We carried out clinical trials by deploying the WE-CARE systems at Peking University Hospital. The clinical results clearly showed that our solution achieves a high detection rate of over 95% against common types of anomalies in ECG, while it only incurs a small detection latency around one second, both of which meet the criteria of real-time medical diagnosis. As demonstrated by the clinical results, the WE-CARE system is a useful and efficient mHealth (mobile health) tool for the cardiovascular disease diagnosis and treatment in medical platforms.

WE-CARE: An Intelligent Mobile Telecardiology System to Enable mHealth Applications

This paper compares the emotional pattern recognition method between standard BP neural network classifier and BP neural network classifier improved by the L-M algorithm. Then we compare the method Support Vector Machine (SVM) to them. Experiment analyzes wavelet transform of surface Electromyography (EMG) to extract the maximum and minimum wavelet coefficients of multi-scale firstly. We then input the two kinds of classifier of the structural feature vector for emotion recognition. The experimental result shows that the standard BP neural network classifier, L-M improved BP neural network classifier and support vector machine's overall pattern recognition rate is 62.5%, 83.33% and 91.67 respectively. Experimental result shows that feature vector extracted by the wavelet transform can characterize emotional patterns through the comparison with the BP neural network classifier and Support Vector Machine, indicating that the Support Vector Machine have a stronger emotional recognition effect.

/pdf/emotion-recognition-of-emg-based-on-improved-l-m-bp-neural-1b0g3q4f53.pdf

Emotion Recognition of EMG Based on Improved L-M BP Neural Network and SVM

Cardiovascular disease (CVD), principally heart disease and stroke, has become the worldwide leading killer for people of all racial and ethnic groups. In China, this disease contributes 41% to all deaths each year, and it costs the country more than 400 billion US dollars in last decade, including health expenditures and lost productivity, which continues to grow as the population size increases. Recent research findings indicate that CVD can be effectively prevented to some extent by an interdisciplinary approach that combines ICT (Information and Communication Technology) and healthcare applications. Based on this motivation, we developed WE-CARE, a Wearable Efficient teleCARdiology systEm using mobile 7-lead ECG devices. Our WE-CARE system surpasses existing resting ECG systems that reside in hospitals owing to its improved mobility brought by wearable and mobile ECG devices. Meanwhile, it outperforms those conventional dynamic 1-lead or 3-lead ECG systems for mobile users in terms of the appropriate ECG information required for clinical proposes. We carried out clinical trials by deploying the WE-CARE systems at People Hospital of Peking University. The results showed that our solution achieves a high detection rate over 95% against common types of anomalies in ECG signals, while its risk alert delay is limited around one second, both of which match the criteria of real-time healthcare monitoring very well. As demonstrated by the results, the WE-CARE system is a useful and efficient tool for the cardiovascular disease prevention and daily monitoring.

WE-CARE: A wearable efficient telecardiology system using mobile 7-lead ECG devices

Eyes location is an important basis for a face recognition system. In this paper, we proposed a novel method to improve the eyes location accuracy in the face image, which detected eyes by combining dual-orientation Gabor filters and eye templates (left and right) , which were got from the average of 20 eyes images collected randomly from the face database, and the eyes. First, YCrCb color space was used to extract face region from image, and then get the interest regions of human eyes by dual-orientation Gabor filters. Finally, the interest regions were matched with the eye templates and the position of eyes were located accurately. The experimental results on images of FERET database indicate that the proposed method was able to achieve a location accuracy of 83% in 1477*2 eyes, and it also effective in video images. Using this method might enhance the performance of face detection and recognition in the video surveillance system. Keywords-Eyes Location; Dual-orientation Gabor Filters; Eye Templates

Eyes Location Based on Dual-Orientation Gabor Filters and Templates

Forehead sEMG signal based HMI for hands-free control

The ratio of signal-noise of the ECG(electrocardiogram) signal is low.In order to eliminate the noise in ECG,improve the performance of ECG monitors and the diagnosis efficiency of computer-aided system,various kinds of noise rejection methods have been proposed.Generally,maybe all of them are useful to the denoising in some certain,but not full.In this thesis,a series of denoising steps are introduced.As a result,a relatively original ECG waveform is detected.

Integrated denoising steps for electrocardiogram signal

A novel method is presented based on the multi-feature information fusion to identify hand motion commands for controlling a prosthetic hand.The integral absolute value,autoregressive(AR) model coefficients,and linear cepstrum coefficients are extracted as feature parameters from time segments of the surface myoelectric signals.Back-propagation(BP) neural networks are used as classification models in various feature domains,and then outputs of these networks are combined by the D-S evidence theory to obtain a consensus classification result.Experimental results show that the recognition rate of the proposed classification strategy is higher than that of the single feature domain method,and the strategy is more efficient than the conventional structural neural network.

Surface Myoelectric Signal Information Fusion and Motion Classification

The application of improved BP neural network together with the wavelet transform to the classification of surface EMG signal is described. The data reduction and preprocessing of the signal are performed by wavelet transform. The network can identify such four kinds of forearm movements with a high accuracy as hand extension, clench fist, forearm pronation and forearm supination. This paper compares the results by standard BP algorithm with that of Bayesian regularization together with LM algorithm. Experimental result shows that the improved BP neural network has a great potential when applied to electromechanical prosthesis control because of its enhanced training speed and identification accuracy.

Application of improved BP algorithm to surface EMG signal classification

As the limitation in the precision and speed of template matching,a modified eye locating method for template matching with extend templates selected by correlation coefficient algorithm is proposed.At the same time,this paper utilizes the rough location and accurate location to speed up the process.The inaccurate location due to the high similarity between the eyes and brows is avoided by changing the eye template to eye-brow template.The algorithm reduces the redundancy of the template set by taking the correlation coefficient to select the efficient templates,and then,the man-made error in template select is avoided.Thus the location precision and speed are improved obviously.

Modified template matching based algorithm for eye location

Based on decision tree analysis, this paper proposes an approach of FO contours prediction for natural speech. In this approach, syllabic maximum pitch, minimum pitch and duration, as well as a set of normalized template of syllabic FO shape are used to generate FO contours for each syllable. To maintain the pitch continuity on sentence level, the predicted pitch information of previous syllable is combined with context parameters to predict the pitch contour of current syllable. In testing data, the syllabic RMS error between real pitch and predicted pitch is less than 10 Hz. This fact demonstrates the continuity of pitch contours and high prediction accuracy. Finally, it achieves natural sounding prosody when applying this method to a PSOLA speech synthesis.

Wang Zhi-zhong

Papers

Integrated denoising steps for electrocardiogram signal

Surface Myoelectric Signal Information Fusion and Motion Classification

Application of improved BP algorithm to surface EMG signal classification

Modified template matching based algorithm for eye location

Predict FO contours prediction on sentence level