A Novel Optimized Deep Learner with Lightweighted Encryption-Based Speaker Verification

In the modern digitalized world, Speaker verification (SV) system is essential for authorizing the client’s credentials. To design an effective SV system, MGWOVSW-CAES-GMM system has been proposed. In this system, the Modified Grey Wolf Optimization (MGWO) technique was employed to optimize the variable sliding window size, FMPM features and training variables. The optimized features were watermarked and encrypted using a Chaotic-based Advanced Encryption Standard (CAES). Once the encryption process was completed, the encrypted features were forwarded to the recipient who executes the decryption and de-watermarking processes. At last, the decrypted features were classified using Gaussian Mixture Model (GMM) classifier. Conversely, MGWO has poor convergence rate and ineffective searching results. Hence, this article proposes an EEHOVSW-CAES-GMM system in which Enhanced Elephant Herding Optimization (EEHO) algorithm is applied instead of MGWO. On the contrary, the computational complexity of GMM classifier is high and its efficiency is less while increasing the number of features. For this reason, a Deep Neural Network (DNN) classifier is employed instead of GMM for recognizing the decrypted features and authorize the speaker’s identity. Besides, the parameters utilized in DNN topology are optimized using two different systems such as MGWOVSW-CAES-DNN and EEHOVSW-CAES-DNN for reducing the computational complexity and increasing the classification accuracy effectively when using more number of features. By using these classifiers, the speaker’s identity is verified and the attacks during the transmission are prevented with the highest security level.

GMM-Based Evaluation of Synthetic Speech Quality Using 2D Classification in Pleasure-Arousal Scale

The paper focuses on the description of a system for the automatic evaluation of synthetic speech quality based on the Gaussian mixture model (GMM) classifier. The speech material originating from a real speaker is compared with synthesized material to determine similarities or differences between them. The final evaluation order is determined by distances in the Pleasure-Arousal (P-A) space between the original and synthetic speech using different synthesis and/or prosody manipulation methods implemented in the Czech text-to-speech system. The GMM models for continual 2D detection of P-A classes are trained using the sound/speech material from the databases without any relation to the original speech or the synthesized sentences. Preliminary and auxiliary analyses show a substantial influence of the number of mixtures, the number and type of the speech features used the size of the processed speech material, as well as the type of the database used for the creation of the GMMs on the P-A classification process and on the final evaluation result. The main evaluation experiments confirm the functionality of the system developed. The objective evaluation results obtained are principally correlated with the subjective ratings of human evaluators; however, partial differences were indicated, so a subsequent detailed investigation must be performed.

Speaker Recognition System Based on Wavelet Features and Gaussian Mixture Models

Identification of a person’s voice from the different voices is known as speaker recognition. The speech signals of individuals are selected by means of speaker recognition or identification. In this work, an efficient method for speaker recognition is made by using Discrete Wavelet Transform (DWT) features and Gaussian Mixture Models (GMM) for classification is presented. The input speech signal features are decomposed by DWT into subband coefficients. The DWT subband coefficient features are the input for the classification. Classification is made by GMM classifier at 4, 8, 16 and 32 Gaussian component levels. Results show a better accuracy of 96.18% speaker signals using DWT features and GMM classifier

Development of Quranic Reciter Identification System using MFCC and GMM Classifier

Nowadays, there are many beautiful recitation of Al-Quran available. Quranic recitation has its own characteristics, and the problem to identify the reciter is similar to the speaker recognition/identification problem. The objective of this paper is to develop Quran reciter identification system using Mel-frequency Cepstral Coefficient (MFCC) and Gaussian Mixture Model (GMM). In this paper, a database of five Quranic reciters is developed and used in training and testing phases. We carefully randomized the database from various surah in the Quran so that the proposed system will not prone to the recited verses but only to the reciter. Around 15 Quranic audio samples from 5 reciters were collected and randomized, in which 10 samples were used for training the GMM and 5 samples were used for testing. Results showed that our proposed system has 100% recognition rate for the five reciters tested. Even when tested with unknown samples, the proposed system is able to reject it.

Abstract 17684: A Novel System for the Rapid and Automated Detection of Atrial Fibrillation

Introduction: Home telemetry monitoring with accurate automated rhythm classification can have important clinical benefits in the timely diagnosis and appropriate management of patients with atrial fibrillation (AF). We clinically validated a novel personal e-Health device and algorithm developed to distinguish AF from sinus rhythm (SR). Methods: A handheld electrocardiogram (ECG) recording system (Maestro) and signal processing platform were developed. The Maestro provides an LCD interface that continuously shows an ECG, heart rate, and heart rhythm status. Twenty second ECG signals analogous to Lead I were acquired from 66 patients presenting to the arrhythmia clinic at the University of Michigan Hospital either in SR or AF. Electrograms were segmented into non-overlapping 6-second samples and one random segment per patient was selected for analysis by the Maestro system. Simultaneous 5 or 12 lead ECGs were obtained from these patients and 3 expert physicians blinded to the Maestro analysis identified the rhythm as SR or AF. The Maestro system applied several signal conditioning algorithms to each ECG sample. The dimensionless temporal R-R interval variability (VRR) index and spectral frequency dispersion metric (FDM) were computed. Results: The 2-dimensional scatter-gram of the samples demonstrated 2 distinct clusters of VRR and FDM for patients with SR and AF. The VRR index clusters for SR and AF patients were 0.018 ± 0.013 and 0.187 ± 0.073 (mean±std), respectively (p < 0.001). The FDM clusters for SR and AF patients occurred at 10.5 ± 5.916 and 15.892 ± 3.337, respectively (p < 0.001). We developed a Gaussian Mixed Model (GMM) classifier to distinguish between the AF and SR clusters. Only after the GMM classifier was obtained were the Maestro classifications compared to the physicians’ readings. The algorithm correctly categorized AF (N = 46) and SR (N = 20) for all Maestro segments analyzed with 100% specificity and sensitivity. Conclusion: The Maestro handheld telemetry unit utilizes a novel classification algorithm and was demonstrated to acquire and automatically analyze 6-second electrograms for rapid and accurate classification of patients in SR or AF in this initial clinical validation trial.