Generative adversarial networks for anonymous acneic face dataset generation.

It is well known that the performance of any classification model is effective if the dataset used for the training process and the test process satisfy some specific requirements. In other words, the more the dataset size is large, balanced, and representative, the more one can trust the proposed model's effectiveness and, consequently, the obtained results. Unfortunately, large-size anonymous datasets are generally not publicly available in biomedical applications, especially those dealing with pathological human face images. This concern makes using deep-learning-based approaches challenging to deploy and difficult to reproduce or verify some published results. In this paper, we propose an efficient method to generate a realistic anonymous synthetic dataset of human faces, focusing on attributes related to acne disorders at three distinct levels of severity (Mild, Moderate, and Severe). Notably, our approach initiates from a small dataset of facial acne images, leveraging generative techniques to augment and diversify the dataset, ensuring comprehensive coverage of acne severity levels while maintaining anonymity and realism in the synthetic data. Therefore, a specific hierarchy StyleGAN-based algorithm trained at distinct levels is considered. Moreover, the utilization of generative adversarial networks for augmentation offers a means to circumvent potential privacy or legal concerns associated with acquiring medical datasets. This is attributed to the synthetic nature of the generated data, where no actual subjects are present, thereby ensuring compliance with privacy regulations and legal considerations. To evaluate the performance of the proposed scheme, we consider a CNN-based classification system, trained using the generated synthetic acneic face images and tested using authentic face images. Consequently, we show that an accuracy of 97.6% is achieved using InceptionResNetv2. As a result, this work allows the scientific community to employ the generated synthetic dataset for any data processing application without restrictions on legal or ethical concerns. Moreover, this approach can also be extended to other applications requiring the generation of synthetic medical images.

Machine-Learning-Based Muscle Control of a 3D-Printed Bionic Arm.

In this paper, a customizable wearable 3D-printed bionic arm is designed, fabricated, and optimized for a right arm amputee. An experimental test has been conducted for the user, where control of the artificial bionic hand is accomplished successfully using surface electromyography (sEMG) signals acquired by a multi-channel wearable armband. The 3D-printed bionic arm was designed for the low cost of 295 USD, and was lightweight at 428 g. To facilitate a generic control of the bionic arm, sEMG data were collected for a set of gestures (fist, spread fingers, wave-in, wave-out) from a wide range of participants. The collected data were processed and features related to the gestures were extracted for the purpose of training a classifier. In this study, several classifiers based on neural networks, support vector machine, and decision trees were constructed, trained, and statistically compared. The support vector machine classifier was found to exhibit an 89.93% success rate. Real-time testing of the bionic arm with the optimum classifier is demonstrated.

Methamphetamine drug abuse and addiction: Effects on face asymmetry.

The abnormal aging mechanism associated with drug abuse results in poor performance of face recognition systems on illicit drug addicts (mainly methamphetamine). Consequently, the high correlation between drug addiction and crime exaggerates the urge for further investigations to originate and overcome this problem. Concurrently, face asymmetry was found to play a significant role in face recognition and age estimation. Therefore, facial asymmetry assessment for meth-addicts is highly serviceable, acknowledging how meth addiction accelerates biological aging and causes severe face distortion. In this work, we address facial asymmetry for meth-addicts compared with ordinary people. We assess facial asymmetry by employing the most credible state-of-the-art tools for local and global two-dimensional (2D) methods. More specifically, we use a classical bilateral-based metric for local analysis, combined with a proposed global approach, that we refer to as the Area Mismatch metric, to give a vivid overview of geometrical facial asymmetry. Finally, we construct a metric for textural facial asymmetry assessment by employing the Structural Similarity Index (SSIM) for dual regions in a given face. We apply the aforementioned metrics on two databases, a recently collected meth-addicted database and a regular aging database (FERET). Statistical analysis indicated a significant increment of facial asymmetry for meth addicts while aging, three to five times more than ordinary people. This study definitively answers the question regarding the correlation between meth abuse and addiction and the increase of facial asymmetry. Also, it confirms previous findings concerning aging and increased facial asymmetry.

A novel biometric authentication approach using ECG and EMG signals.

Security biometrics is a secure alternative to traditional methods of identity verification of individuals, such as authentication systems based on user name and password. Recently, it has been found that the electrocardiogram (ECG) signal formed by five successive waves (P, Q, R, S and T) is unique to each individual. In fact, better than any other biometrics' measures, it delivers proof of subject's being alive as extra information which other biometrics cannot deliver. The main purpose of this work is to present a low-cost method for online acquisition and processing of ECG signals for person authentication and to study the possibility of providing additional information and retrieve personal data from an electrocardiogram signal to yield a reliable decision. This study explores the effectiveness of a novel biometric system resulting from the fusion of information and knowledge provided by ECG and EMG (Electromyogram) physiological recordings. It is shown that biometrics based on these ECG/EMG signals offers a novel way to robustly authenticate subjects. Five ECG databases (MIT-BIH, ST-T, NSR, PTB and ECG-ID) and several ECG signals collected in-house from volunteers were exploited. A palm-based ECG biometric system was developed where the signals are collected from the palm of the subject through a minimally intrusive one-lead ECG set-up. A total of 3750 ECG beats were used in this work. Feature extraction was performed on ECG signals using Fourier descriptors (spectral coefficients). Optimum-Path Forest classifier was used to calculate the degree of similarity between individuals. The obtained results from the proposed approach look promising for individuals' authentication.

A novel scheme for joint multi-channel ECG-ultrasound image compression.

In this paper, we introduce a novel approach to compress jointly a Multi-Channel Electrocardiogram (MCE) and an ultrasound image. We will show that this technique allows better performances, in terms of compression ratio (CR) compared to coding separately both modalities. In this approach, scaled ECG samples are inserted within the high frequencies of the ultrasound image after its decomposition on wavelet basis. The new standard JPEG2000 is then applied on the packed data for both coding and decoding purpose. Finally, the reconstruction quality is evaluated using the PSNR (Peak Signal Noise Ratio) and the PRD (Percent Root Mean Square Difference), respectively for both the ultrasound image and the ECG signals.

Parametrical modelling of a premature ventricular contraction ECG beat: comparison with the normal case.

The aim of this paper is to analyse a parametrical Gaussian kernel based model. The proposed model is tested on two types of electrocardiogram (ECG) beats, the normal case beat and the premature ventricular contraction (PVC) one. Basically, the model is constituted of N Gaussians where their corresponding parameters are estimated by optimising a specific criterion. The modelling technique has been validated using MIT/BIH databases. As a result of this study, we show that a normal beat can be modelled using 18 parameters and only 15 parameters are needed to reconstruct the PVC one.

Evaluation of a none compressed 12-lead ECG transmission via internet using IPSec.

Electrocardiogram (ECG) transmission in real time could involve some problems with the quality of services and security. This paper aims to evaluate a none compressed 12-lead ECG transmission using the Secure Internet Protocol (IPSec), compared with the popular Internet Protocol (IP). Using an analytical model, the transmission performance is estimated in terms of end-to-end delay and loss rate. Our results show that ECG transmission could be assured both security and quality of services.

A new vision on the averaging technique for the estimation of non-stationary Brainstem Auditory-Evoked Potentials: application of a metaheuristic method.

The aim of this paper consists in highlighting the use of the averaging technique in some biomedical applications, such as evoked potentials (EP) extraction. We show that this technique, which is generally considered as classical, can be very efficient if the dynamic model of the signal to be estimated is a priori known. Therefore, using an appropriate model and under some specific conditions, one can show that the estimation can be performed efficiently even in case of a very low signal to noise ratio (SNR), which occurs when handling Brainstem Auditory-Evoked Potentials.

ECG compression by modelling the instantaneous module/phase of its DCT.

Recent developments in compression methods on the non-linear and non-stationary data, such as electrocardiograms (ECG), have received large attention by the time-frequency analysts. The technique presented in this paper is based on parametrical modeling the instantaneous module as well as the instantaneous phase, estimated directly from the Discrete Cosine Transform (DCT) of each ECG beat. The estimated parameters are then used to reconstruct each recorded beat. In order to evaluate the performance of our technique, data recorded from the MIT-BIH arrhythmia database are used.

Fast simulated annealing algorithm for BAEP time delay estimation using a reduced order dynamic model.

The extraction of Brainstem Auditory Evoked Potentials (BAEPs), from the EEG background, is of high clinical interest. The present paper can be considered as a continuation of our previous work related to the BAEP estimation in endocochlear pathologies. In our previous published work, we proposed a technique for time delay estimation on the basis of the BAEPs in response to successive stimulations. Put in a different manner, our objective was the estimation of the dynamics of the cochlea that is responsible for the delayed responses. The estimation technique was based on optimization of a non-linear criterion by means of Simulated Annealing Time Delay Estimation (SATDE) algorithm. However, it is well known that such heuristic algorithms are time consuming and largely depend on the number of parameters to be estimated. The present paper demonstrates that modeling the non-stationarity of responses considerably decreases the convergence time to the global minimum. The newly proposed method in this paper, called Fast Simulated Annealing Time Delay Estimation (FSATDE) algorithm, has been validated on both simulated and real signals.

An input-delay neural-network-based approach for piecewise ECG signal compression.

We propose an input delay neural network (IDNN) based time series prediction algorithm for compressing electrocardiogram (ECG) signals. Our algorithm has been tested and successfully compared vis-à-vis other popular techniques for its compression efficiency and reconstruction capability.