Maen Takruri

Toward improved control of prosthetic fingers using surface electromyogram (EMG) signals

A fundamental component of many modern prostheses is the myoelectric control system, which uses the electromyogram (EMG) signals from an individuals muscles to control the prosthesis movements. Despite the extensive research focus on the myoelectric control of arm and gross hand movements, more dexterous individual and combined fingers control has not received the same attention. The main contribution of this paper is an investigation into accurately discriminating between individual and combined fingers movements using surface EMG signals, so that different finger postures of a prosthetic hand can be controlled in response. For this purpose, two EMG electrodes located on the human forearm are utilized to collect the EMG data from eight participants. Various feature sets are extracted and projected in a manner that ensures maximum separation between the finger movements and then fed to two different classifiers. The second contribution is the use of a Bayesian data fusion postprocessing approach to maximize the probability of correct classification of the EMG data belonging to different movements. Practical results and statistical significance tests prove the feasibility of the proposed approach with an average classification accuracy of ~90% across different subjects proving the significance of the proposed fusion scheme in finger movement classification.

Towards limb position invariant myoelectric pattern recognition using time-dependent spectral features.

Recent studies in Electromyogram (EMG) pattern recognition reveal a gap between research findings and a viable clinical implementation of myoelectric control strategies. One of the important factors contributing to the limited performance of such controllers in practice is the variation in the limb position associated with normal use as it results in different EMG patterns for the same movements when carried out at different positions. However, the end goal of the myoelectric control scheme is to allow amputees to control their prosthetics in an intuitive and accurate manner regardless of the limb position at which the movement is initiated. In an attempt to reduce the impact of limb position on EMG pattern recognition, this paper proposes a new feature extraction method that extracts a set of power spectrum characteristics directly from the time-domain. The end goal is to form a set of features invariant to limb position. Specifically, the proposed method estimates the spectral moments, spectral sparsity, spectral flux, irregularity factor, and signals power spectrum correlation. This is achieved through using Fourier transform properties to form invariants to amplification, translation and signal scaling, providing an efficient and accurate representation of the underlying EMG activity. Additionally, due to the inherent temporal structure of the EMG signal, the proposed method is applied on the global segments of EMG data as well as the sliced segments using multiple overlapped windows. The performance of the proposed features is tested on EMG data collected from eleven subjects, while implementing eight classes of movements, each at five different limb positions. Practical results indicate that the proposed feature set can achieve significant reduction in classification error rates, in comparison to other methods, with ≈8% error on average across all subjects and limb positions. A real-time implementation and demonstration is also provided and made available as a video supplement (see Appendix A).

Drift aware wireless sensor networks

The focus of wireless sensor networks is to develop low cost sensors with sufficient computing and communication capabilities to support networked sensing applications. The emphasis on lower cost led to sensors that are less accurate and less reliable than their wired sensor counterparts. Sensors usually suffer from both random and systematic (bias) problems. Even when the sensors are properly calibrated at the time of their deployment, they develop drift in their readings leading to biased sensor measurements. The drift in this context is defined as a unidirectional long-term change in the sensor measurement. Assuming that neighboring sensors have correlated measurements and noting that the instantiation of drift in a sensor is uncorrelated with other sensors and inspired by the resemblance of registration problem in radar target tracking with the bias error problem in sensor networks we devise a novel algorithm for detecting and correcting sensors drifts and show how it improves the reliability and the effective life of the network.

Distributed Recursive Algorithm for Auto Calibration in Drift Aware Wireless Sensor Networks

The purpose for wireless sensor networks is to deploy low cost sensors with sufficient computing and communication capabilities to support networked sensing applications. The emphasis on lower cost led to sensors that are less accurate and less reliable than their wired sensor counterparts. Sensors usually suffer from both random and systematic bias problems. Even when the sensors are properly calibrated at the time of their deployment, they develop drift in their readings leading to biased sensor measurements. The drift in this context is defined as a unidirectional long-term change in the sensor measurement. We assume that neighboring sensors have correlated measurements and that the instantiation of drift in a sensor is uncorrelated with other sensors. As an extension of our results in [1], and inspired by the resemblance of registration problem in radar target tracking, we propose a distributed recursive Bayesian algorithm for auto calibration of wireless sensors in the presence of slowly varying drifts. The algorithm detects and corrects sensor drifts and improves the reliability and the effective life of the network.

The automatic identification of melanoma by wavelet and curvelet analysis: Study based on neural network classification

This paper proposes an automatic skin cancer (melanoma) classification system. The input for the prosed system is a collected data images, it followed by different image processing procedures to enhance the image properties. Two segmentation methods used to identify the normal skin cancer from malignant skin and to extract the useful information from these images that passed to the classifier for training and testing. The features used for classification is the coefficients created by Wavelet decompositions and simple wrapper curvelet. Curvelet is suitable for the image that contains oriented texture and cartoon edges. Recognition accuracy of the three layers back-propagation neural network classifier with wavelet is 51.1% and with curvelet is 75. 6% in digital images database.

Recursive Bayesian Approaches for Auto Calibration in Drift Aware Wireless Sensor Networks

The purpose for wireless sensor networks is to deploy low cost sensors with sufficient computing and communication capabilities to support networked sensing applications. Even when the sensors are properly calibrated at the time of their deployment, they develop drift in their readings leading to biased sensor measurements. Noting that a physical phenomenon in a certain area follows some spatio-temporal correlation, we assume that the sensors readings in that area are correlated. We also assume that the instantiations of drifts are uncorrelated. Based on these assumptions, and inspired by the resemblance of registration problem in radar target tracking with the bias error problem in wireless sensor networks, we follow a Bayesian framework to solve the Drift/Bias problem in wireless sensor networks. We present two methods for solving the drift problem in a densely deployed sensor network, one for smooth drifts and the other for unsmooth drifts. We also show that both methods successfully detect and correct sensor errors and extend the effective life time of the sensor network.

Auto calibration in drift aware wireless sensor networks using the interacting multiple model algorithm

The purpose for wireless sensor networks is to deploy low cost sensors with sufficient computing and communication capabilities to support networked sensing applications. The emphasis on lower cost led to sensors that are less accurate and less reliable than their wired sensor counterparts. Sensors usually suffer from both random and systematic bias problems. Even when the sensors are properly calibrated at the time of their deployment, they develop drift in their readings leading to biased sensor measurements. The drift in this context is defined as a unidirectional long-term change in the sensor measurement. Assuming that neighboring sensors have correlated measurements and noting that the instantiation of drift in a sensor is uncorrelated with other sensors, we present the methodology for detecting and correcting sensors smooth and steep drifts. The methodology improves the reliability and the effective life of the network.

Automatic non-invasive recognition of melanoma using Support Vector Machines

This paper proposes an automated non-invasive system for skin cancer (melanoma) detection based on Support Vector Machine classification. The proposed system uses a number of features extracted from the Grey Level Co-occurrence Matrices (GLCM) of the gray-scale skin lesion images, and color features obtained from the original color images. The dataset used include digital images for skin lesions that are either benign or malignant. The purpose of this project is to provide a system that can classify digital images of skin lesions as benign or malignant (Melanoma). The testing accuracy obtained by the Support Vector Machine classifier used in this experiment is 82.7% for the GLCM features and 81.48% for both GLCM and Color features using ROI segmentation. The proposed system has resulted in a sensitivity of 83.6 % for the case of GLCM and 83.33% for the case of GLCM and color using ROI segmentation. It has also resulted in a specificity of 80% for the case of GLCM and 76.19% for the case of both GLCM and Color features using ROI segmentation. The obtained sensitivity and specificity results are comparable to those obtained by dermatologists. Consequently, this can increase the chance of the survival from Melanoma.

Automatic recognition of melanoma using Support Vector Machines: A study based on Wavelet, Curvelet and color features

This paper proposes an automated non-invasive system for skin cancer (melanoma) detection based on Support Vector Machine classification. The proposed system uses a number of features extracted from the Wavelet or the Curvelet decomposition of the grayscale skin lesion images and color features obtained from the original color images. The dataset used include both digital images and Dermoscopy images for skin lesions that are either benign or malignant. The recognition accuracy obtained by the Support Vector Machine classifier used in this experiment is 87.7.1% for the Wavelet based features and 83.6. 6% for the Curvelet based ones. The proposed system also resulted in a sensitivity of 86.4 % for the case of Wavelet and 76.9% for the case of Curvelet. It also resulted in a specificity of 88.1% for the case of Wavelet and 85.4% for the case of Curvelet. The obtained sensitivity and specificity results are comparable to those obtained by Dermatologists.

Design of decentralized street LED light dimming system

To reduce power consumption due to street lighting, we propose replacing conventional power consuming High Pressure Sodium Lamps and metal Halide lamps with LED lamps which consume much less power. We also propose employing LED light dimming by modifying the light intensity based on the traffic conditions on the road. Most of the dimming systems currently deployed on the streets use computerized and remotely controlled wireless monitoring systems which suffer from complexity, high cost and a response that is dependent on the network data transfer speed. Unlike these systems, we propose, in this paper, a decentralized street LED light dimming system that is installed on each pole and whose dimming action is controlled by the dimming circuit of the pole itself. This yields faster and more reliable response. In addition, the proposed system does not need any additional infrastructure. It is scalable in the sense that it can be used in small or main streets with any number of poles, and flexible in system specifications based on the selection of the coverage distance of the designed motion detector. Simulations show the effectiveness of the proposed system and in saving energy.