Mel Frequency Cepstral Coefficients Features Research Articles

Emotion Recognition in Kurdish Speech from the Sorani Dialect Corpus

Given the increasing need for interactive human-computer applications, the field of employing machine learning algorithms to discern emotions from speech has seen a substantial surge in interest. While emotion recognition systems have made substantial progress in languages like German, English, Spanish, Dutch, and Danish, the availability of comprehensive datasets for the Kurdish language remains notably limited. This paper addresses this gap by focusing on emotion recognition in Sorani Kurdish dialect speech data, which was carefully gathered from openly available videos from the YouTube platform and categorized into four clear supposed emotions: neutral, sadness, happiness, and anger. The study applied both natural Mel Spectrogram and Mel-Frequency Cepstral Coefficient (MFCC) features for various spectrals, followed by the classification models K-Nearest Neighbor (KNN), Multi-Layer Perceptron (MLP), and Support Vector Machine (SVM) to evaluate the results. By closely examining and contrasting the results of using several methods for feature extraction, it was found that SVM obtained a higher accuracy, reaching as much as 85.57%. This is so much more than the first Kurdish emotion classification technique for the recognition of the emotion of the words.

Sound Quality Prediction Method of Dual-Phase Hy-Vo Chain Transmission System Based on MFCC-CNN and Fuzzy Generation

redict the sound quality of dual-phase Hy-Vo chain transmission system noise using a small sample size. Noise acquisition tests are conducted under various working conditions, followed by subjective evaluations using the equal interval direct one-dimensional method. Objective evaluations are performed using the Mel-frequency cepstral coefficient (MFCC). To understand the impact of the MFCC order and the frame number on prediction accuracy, MFCC feature maps of different specifications are analyzed. The dataset is expanded threefold using fuzzy generation with an appropriate membership degree. The convolutional neural network (CNN) is developed, utilizing MFCC feature maps as inputs and evaluation scores as outputs. Results indicate a positive correlation between the frame number and prediction accuracy, whereas higher MFCC orders introduce redundancy, reducing accuracy. The proposed CNN method outperforms three traditional machine learning approaches, demonstrating superior accuracy and resistance to overfitting.

A speech-based convolutional neural network for human body posture classification

Abstract Problem In recent years, computing and sensing advances have helped to develop efficient human posture classification systems, which assist in creating health systems that contribute to enhancing elder’s and disable’s life quality in context-aware and ambient assistive living applications. Other applications of body posture classification include the simulation of human bodies into virtual characters, security applications such as kidnapping and the body position of the kidnapper or victim, which can provide useful information to the negotiator or the rescue team, and other sports applications. Aim This work aims to propose a body posture classification system based on speech using deep learning techniques. Methods All samples pass through the preprocessing phase. The Mel-frequency cepstral coefficient (MFCC) with 12 coefficients was used as features, and the best features were selected by the correlation-based feature selection method. Two supervised learning techniques called artificial hydrocarbon networks (AHN) and convolutional neural networks (CNN) were deployed to efficiently classify body postures and movements. This research aims to detect six human positions through speech, which include walking downstairs, sitting, walking upstairs, running, laying, and walking. The dataset used in this work was prepared by the authors, named the human activity speech dataset. Results The best performance gained was with the AHN classifier of 68.9% accuracy and 67.1% accuracy with the CNN. Conclusion It was concluded that the MFCC features are the most powerful features in body position classification, and the deep learning methods are powerful in classifying body positions.

Classification of Infant Crying Sounds Using SE-ResNet-Transformer

Recently, emotion analysis has played an important role in the field of artificial intelligence, particularly in the study of speech emotion analysis, which can help understand one of the most direct ways of human emotional communication—speech. This study focuses on the emotion analysis of infant crying. Within cries lies a variety of information, including hunger, pain, and discomfort. This paper proposes an improved classification model using ResNet and transformer. It utilizes modified Mel-frequency cepstral coefficient Mel-frequency cepstral coefficient (MFCC) features obtained through feature engineering from infant cries and integrates SE attention mechanism modules into residual blocks to enhance the model’s ability to adjust channel weights. The proposed method achieved 93% accuracy rate in experiments, offering advantages of shorter training time and higher accuracy compared to other traditional models. It provides an efficient and stable solution for infant cry classification.

Data augmentation using a 1D-CNN model with MFCC/MFMC features for speech emotion recognition

Speech emotion recognition (SER) is attractive in several domains, such as automated translation, call centres, intelligent healthcare, and human–computer interaction. Deep learning models for emotion identification need considerable labelled data, which is only sometimes available in the SER industry. A database needs enough speech samples, good features, and a better classifier to identify emotions efficiently. This study uses data augmentation to enhance the amount of input voice samples and address the data shortage issue. The database capacity increases by adding white noise to the speech signals by data augmentation. In this work, the Mel-frequency Cepstral Coefficient (MFCC) and Mel-frequency Magnitude Coefficient (MFMC) features, along with a one-dimensional convolutional neural network (1D-CNN), are used to classify speech emotions. The datasets utilized to estimate the model's enactment were AESDD, CAFE, EmoDB, IEMOCAP, and MESD. The data augmentation with the 1D-CNN (MFMC) model performed best, with an average accuracy of 99.2% for AESDD, 99.5% for CAFE, 97.5% for EmoDB, 92.4% for IEMOCAP and 96.9% for the MESD database. The proposed 1D-CNN (MFMC) with data augmentation outperforms the 1D-CNN (MFCC) without data augmentation in emotion recognition.

Audio Signal Recognition in Complex Environments Using Sparse Representation

Recognizing audio signals in complex environments is crucial for acquiring adequate information. This paper integrates the sparse expression algorithm with Mel-frequency cepstral coefficient (MFCC) features. The combined approach was applied in convolutional neural network classifiers to recognize acoustic scenes in audio signals within complex environments. The algorithm was then simulated and tested using the TUT Sound Events 2016 and TUT Acoustic Scenes 2016 datasets. In the experiments, the efficacy of the developed sparse feature extraction method was validated. Then, the appropriate sparse dictionary size was determined. The algorithm was subsequently compared with two recognition algorithms based on sparse and MFCC features, respectively. It was found that the extraction approach proposed in this paper had a higher signal-to-noise ratio. The results revealed variations in the required sparse dictionary size for different datasets: 75 for TUT Sound Events 2016 and 150 for TUT Acoustic Scenes 2016. The MFCC-combined recognition algorithm demonstrated the fastest convergence during training among the three audio scene recognition algorithms. For both the TUT Sound Events 2016 and TUT Acoustic Scenes 2016 datasets, the MFCC-combined recognition algorithm achieved the highest classification accuracy, and the recognition accuracy for the former dataset was higher.

A Multi-Feature Fusion Approach for Dialect Identification using 1D CNN

The phonological variety of Kurdish, a language with several dialects, poses a distinct problem in automatically identifying dialects. This study examines and evaluates several sound criteria for identifying Kurdish dialects: Badini, Hawrami, and Sorani. We deployed a dataset including 6,000 samples and utilized a mix of 1D convolutional neural networks (CNN) and fully connected layers to conduct the identification job. Our study aimed to assess the efficacy of different sound characteristics in accurately identifying dialects. We employed the Mel-frequency Cepstral Coefficients (MFCC) and other features such as the Mel spectrogram, spectral contrast, and polynomial features to extract the sound characteristics. We conducted training and testing of our models utilizing both individual characteristics and a composite of all features. Our analysis revealed that the identification task achieved excellent accuracy rates, suggesting a promising potential for success. We achieved 95.75% accuracy using MFCC combined with a Mel spectrogram. The accuracy improved by including contrast in the MFCC feature extraction process to 91.42%. Similarly, using poly_features resulted in an accuracy of 90.83%. Remarkably, accuracy reached a maximum of 96.5% when all the attributes were combined.

A multi-modal Parkinson’s disease diagnosis system from EEG signals and online handwritten tasks using grey wolf optimization based deep learning model

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by the gradual deterioration of motor function, affecting speech, writing, muscle control, and mobility. The existing studies have not utilized both the electroencephalography (EEG) signals and online handwritten tasks together to diagnose PD. The studies have also not explored the EEG signals collected from specific brain regions like the substansia niagra (SN) and ventral tegmental area (VTA), crucial for dopamine production linked to PD. This article proposes a multi-modal PD diagnosis system from EEG signals (collected from SN and VTA regions of the brain), collected during performing online handwritten tasks, using grey wolf optimization (GWO) algorithm. Mel-frequency cepstral coefficients (MFCC) features have been generated from the EEG signals and optimized by the GWO algorithm. The classification (diagnosis) experiments on the optimal number of feature values, obtained from GWO algorithm, have been carried out using bidirectional long short-term memory (BLSTM) variant of recurrent neural network (RNN). The classification experiments have also been conducted using support vector machine (SVM), bagged random forest (BRF), and long short-term memory (LSTM) variant of RNN classifier to have a performance comparison with the proposed method. The effectiveness of the introduced PD diagnosis system has been analyzed on a self-generated dataset named EEG signal based on online handwriting (ESOH). A maximum classification accuracy of 99.30% has been achieved from the proposed PD diagnosis system. The experimental outcomes illustrate that the introduced PD diagnosis system outperforms the state-of-the-art PD diagnosis systems relying on the EEG signals for diagnosing the PD.

Application of Deep Learning for Voice Command Classification in Turkish Language

In this study, a deep learning model was developed for the recognition and classification of voice commands using the Turkish Speech Command Dataset. The division of training, validation, and test sets was carried out on an individual basis. This approach aims to prevent the model from memorizing and to enhance its generalization capability. The model was trained using Mel-Frequency Cepstral Coefficients (MFCC) features extracted from voice files, and its classification performance was evaluated in detail. The findings indicate that the model successfully classifies voice commands with a high accuracy rate, achieving an overall accuracy of 92.3% on the test set, highlighting the potential of deep learning approaches in voice recognition technologies

Feature analysis and recognition of fiber breakage AE signals after propagation

Acoustic emission (AE) is a non-destructive testing technique, and establishing correlations between AE signals and material damage modes is one of its primary challenges. However, it is difficult to identify damage modes in ceramic matrix composites (CMCs) due to AE signal attenuation occurring after propagation and complex damage modes. In this study, AE signals generated by the breakage of C and SiC fibers were monitored at different distances and angles on the C/SiC plate. The attenuation of energy and the frequency spectra were analyzed. The Mel-frequency cepstral coefficient (MFCC) method was used to analyze the waveform data of AE signals and extract MFCC features. To identify the damage caused by C and SiC fiber breakage, AE parameter features and MFCC features were selected as inputs, and a fully connected neural network was constructed to train a supervised pattern recognition model. The results show that the MFCC feature has higher recognition accuracy than the traditional feature when AE is used for damage identification.

Implementation of Data Mining for Speech Recognition Classification of Sundanese Dialect Using KNN Method with MFCC Feature Extraction

The importance of preservation and development of speech recognition technology for regional languages such as Sundanese, which have unique phonetic characteristics. Regional language speech recognition can assist in the development of local, educational, and cultural preservation applications to implement and evaluate the effectiveness of the combination of MFCC and KNN methods in classifying Sundanese dialect speech recognition. Methods used include trait extraction with MFCC, which converts voice data into numerical representations based on frequency characteristics, and classification with KNN, which groups data based on similarity to train data. The Dataset used consisted of speech recordings of Western and Southern Sundanese dialects. The results showed that the k-Nearest Neighbors (KNN) method can classify Sundanese dialect speech recognition with an accuracy of 80.00%, showing good ability in distinguishing "Western" and "southern" dialects. Mel-Frequency Cepstral Coefficients (MFCC) proved to be very effective in extracting sound features, helping KNN achieve low error rates. The combination of MFCC and KNN proved effective for speech recognition classification of Sundanese dialects, providing satisfactory results with high accuracy.

DCNN for Pig Vocalization and Non-Vocalization Classification: Evaluate Model Robustness with New Data.

Since pig vocalization is an important indicator of monitoring pig conditions, pig vocalization detection and recognition using deep learning play a crucial role in the management and welfare of modern pig livestock farming. However, collecting pig sound data for deep learning model training takes time and effort. Acknowledging the challenges of collecting pig sound data for model training, this study introduces a deep convolutional neural network (DCNN) architecture for pig vocalization and non-vocalization classification with a real pig farm dataset. Various audio feature extraction methods were evaluated individually to compare the performance differences, including Mel-frequency cepstral coefficients (MFCC), Mel-spectrogram, Chroma, and Tonnetz. This study proposes a novel feature extraction method called Mixed-MMCT to improve the classification accuracy by integrating MFCC, Mel-spectrogram, Chroma, and Tonnetz features. These feature extraction methods were applied to extract relevant features from the pig sound dataset for input into a deep learning network. For the experiment, three datasets were collected from three actual pig farms: Nias, Gimje, and Jeongeup. Each dataset consists of 4000 WAV files (2000 pig vocalization and 2000 pig non-vocalization) with a duration of three seconds. Various audio data augmentation techniques are utilized in the training set to improve the model performance and generalization, including pitch-shifting, time-shifting, time-stretching, and background-noising. In this study, the performance of the predictive deep learning model was assessed using the k-fold cross-validation (k = 5) technique on each dataset. By conducting rigorous experiments, Mixed-MMCT showed superior accuracy on Nias, Gimje, and Jeongeup, with rates of 99.50%, 99.56%, and 99.67%, respectively. Robustness experiments were performed to prove the effectiveness of the model by using two farm datasets as a training set and a farm as a testing set. The average performance of the Mixed-MMCT in terms of accuracy, precision, recall, and F1-score reached rates of 95.67%, 96.25%, 95.68%, and 95.96%, respectively. All results demonstrate that the proposed Mixed-MMCT feature extraction method outperforms other methods regarding pig vocalization and non-vocalization classification in real pig livestock farming.

A lightweight and privacy preserved federated learning ecosystem for analyzing verbal communication emotions in identical and non-identical databases

The lack of vocal emotional expression is a major deficit in social communication disorders. The current scenario of artificial intelligence focuses on collaborative training of deep learning models without losing data privacy. The primary objective of this paper is to propose a federated learning-based classification model to identify and analyze the emotional capabilities of individuals with vocal emotion deficits. The methodology has developed a collaborative and privacy-preserved approach using federated learning for training the deep learning models. The proposed methodology utilizes Mel-frequency Cepstral Coefficients (MFCC) to preprocess audio recordings. The four datasets (RAVDESS, CREMA, TESS, SAVEE) including emotion-based classified audio recordings were collected from open sources. The collected audio recordings are 3 s each and the total data set has 668376 audio files with happy - 175119 files, sad – 172611 files, angry – 176346 files, and normal - 144300 files. Further, the input audio was pre-processed to generate MFCC features. The study began with extracting features from multiple pre-trained DL models as its base model. Then, the performance of the federated learning (FL) model was tested on independent and identically distributed (IID) and non-IID data. Further, this paper presents a federated deep learning-based multimodal system for verbal communication emotions classification that uses audio datasets to meet data privacy requirements by DL on the FL ecosystem. As per the findings, the federated learning trained model provides nearly similar parametric results in comparison to base model training. For IID data, the model had 99.71 % validation accuracy, precision (99.73 %), recall (99.69 %), and validation loss (0.01). The FL architecture with non-IID data outperformed these measures with validation accuracy (99.97 %), precision (99.97 %), recall (99.97 %), and least loss (0). Hence the acquired results support the utilization of federated learning ecosystem-based trained models with identically and non-identically distributed audio features from emotion identification without losing parametric results. In conclusion, the proposed techniques could be applied to identify verbal emotional deficits in individuals and could support developing emerging technological interventions for their well-being.

Crack detection based on mel-frequency cepstral coefficients features using multiple classifiers

Crack detection plays an essential role in evaluating the strength of structures. In recent years, the use of machine learning and deep learning techniques combined with computer vision has emerged to assess the strength of structures and detect cracks. This research aims to use machine learning (ML) to create a crack detection model based on a dataset consisting of 2432 images of different surfaces that were divided into two groups: 70% of the training dataset and 30% of the testing dataset. The Orange3 data mining tool was used to build a crack detection model, where the support vector machine (SVM), gradient boosting (GB), naive Bayes (NB), and artificial neural network (ANN) were trained and verified based on 3 sets of features, mel-frequency cepstral coefficients (MFCC), delta MFCC (DMFCC), and delta-delta MFCC (DDMFCC) were extracted using MATLAB. The experimental results showed the superiority of SVM with a classification accuracy of (100%), while for NB the accuracy reached (93.9%-99.9%), and (99.9%) for ANN, and finally in GB the accuracy reached (99.8%).

Deepfake Audio Detection System

Abstract: To ensure the authenticity of audio material, it is important to establish reliable detection tools that can trace the spread and use of deepfake technologies. This study focuses on the methodology of deepfake audio identification via Mel Frequency Cepstral Coefficients (MFCC) as features with Random Forest as a classifier. By extracting MFCC features from audio clips and using them in a Random Forest model, one can learn unique spectral properties that can help distinguish deepfakes from authentic audio. The Random Forest algorithm, famous for its quality of being able to work well in an ensemble learning paradigm, is utilized to identify patterns that are representative of deepfake manipulation. To ensure the efficiency and reliability of this method, it was tested on a large number of different data sets that included both genuine and fake voice samples. To ensure robustness and generalization, cross-validation techniques are employed, restricting model predictions to the range of 0 to 1 and providing informative error messages for effective diagnosis. Thus, it is an important scientific study helping to develop and strengthen methods for identifying and eliminating threats in the area of artificial sound activity.

Responding to challenge call for machine learning model development in diagnosing respiratory disease sounds

The normal and abnormal sounds from the respiratory system shed great light on medical science by revealing the quality, diseases and changes in people’s lungs. In medicine, this easy and old method, realized by the stethoscope, facilitates the diagnosis of diseases by specialists. This manual method can sometimes lead to wrong decisions in terms of sound detection due to different audibility. Detailed sound analysis is crucial for accurately detecting lung diseases with high mortality rates. As technology advances, the development of automated approaches based on machine learning is of great interest as they provide modern and highly accurate analysis. In today’s most popular topic, i.e., the COVID-19 disaster, the conflict between early detection of respiratory disease and machine learning for sound signal processing is extremely important. In this study, a machine learning model was developed to detect respiratory system sounds such as sneezing/coughing in disease diagnosis. The automatic model and approach development of breath sounds, which carry valuable information, results in early diagnosis and treatment. A successful machine learning model was developed in this study, which was a strong response to the challenge called the “Pfizer digital medicine challenge” on the “OSFHOME” open access platform. In the database provided in this challenge, which consists of 3 parts, features that effectively showed coughing/sneezing sound analysis were extracted from training, testing and validating samples. Based on the Mel frequency cepstral coefficients (MFCC) feature extraction method, mathematical and statistical features were prepared. The sequential forward selection (SFS) feature selection method was used to select the relevant and dominant variables among the obtained features to represent the model fully and accurately. Three different classification techniques were considered for successful respiratory sound classification in the dataset containing more than 3800 sounds. Support vector machine (SVM) with radial basis function (RBF) kernels, decision tree and ensemble aggregation classification methods were used as classification techniques. In an attempt to classify coughing/sneezing sounds from other sounds, SVM with RBF kernels was achieved with 83% success.

A Study on Speech Recognition by a Neural Network Based on English Speech Feature Parameters

In this study, from the perspective of English speech feature parameters, two feature parameters, the mel-frequency cepstral coefficient (MFCC) and filter bank (Fbank), were selected to identify English speech. The algorithms used for recognition employed the classical back-propagation neural network (BPNN), recurrent neural network (RNN), and long short-term memory (LSTM) that were obtained by improving RNN. The three recognition algorithms were compared in the experiments, and the effects of the two feature parameters on the performance of the recognition algorithms were also compared. The LSTM model had the best identification performance among the three neural networks under different experimental environments; the neural network model using the MFCC feature parameter outperformed the neural network using the Fbank feature parameter; the LSTM model had the highest correct rate and the highest speed, while the RNN model ranked second, and the BPNN model ranked worst. The results confirm that the application of the LSTM model in combination with MFCC feature parameter extraction to English speech recognition can achieve higher speech recognition accuracy compared to other neural networks.

Research on sound quality of roller chain transmission system based on multi-source transfer learning

To establish the sound quality evaluation model of roller chain transmission system, we collect the running noise under different working conditions. After the noise samples are preprocessed, a group of experienced testers are organized to evaluate them subjectively. Mel frequency cepstral coefficient (MFCC) of each noise sample is calculated, and the MFCC feature map is used as an objective evaluation. Combining with the subjective and objective evaluation results of the roller chain system noise, we can get the original dataset of its sound quality research. However, the number of high-quality noise samples is relatively small. Based on the sound quality research of various chain transmission systems, a novel method called multi-source transfer learning convolutional neural network (MSTL-CNN) is proposed. By transferring knowledge from multiple source tasks to target task, the difficulty of small sample sound quality prediction is solved. Compared with the problem that single source task transfer learning has too much error on some samples, MSTL-CNN can give full play to the advantages of all transfer learning models. The results also show that the MSTL-CNN proposed in this paper is significantly better than the traditional sound quality evaluation methods.

Automatically detecting OSAHS patients based on transfer learning and model fusion

Objective. Snoring is the most typical symptom of obstructive sleep apnea hypopnea syndrome (OSAHS) that can be used to develop a non-invasive approach for automatically detecting OSAHS patients. Approach. In this work, a model based on transfer learning and model fusion was applied to classify simple snorers and OSAHS patients. Three kinds of basic models were constructed based on pretrained Visual Geometry Group-16 (VGG16), pretrained audio neural networks (PANN), and Mel-frequency cepstral coefficient (MFCC). The XGBoost was used to select features based on feature importance, the majority voting strategy was applied to fuse these basic models and leave-one-subject-out cross validation was used to evaluate the proposed model. Main results. The results show that the fused model embedded with top-5 VGG16 features, top-5 PANN features, and MFCC feature can correctly identify OSAHS patients (AHI > 5) with 100% accuracy. Significance. The proposed fused model provides a good classification performance with lower computational cost and higher robustness that makes detecting OSAHS patients at home possible.

Exploring the Impact of Mismatch Conditions, Noisy Backgrounds, and Speaker Health on Convolutional Autoencoder-Based Speaker Recognition System with Limited Dataset

This paper presents a novel approach to enhance the success rate and accuracy of speaker recognition and identification systems. The methodology involves employing data augmentation techniques to enrich a small dataset with audio recordings from five speakers, covering both male and female voices. Python programming language is utilized for data processing, and a convolutional autoencoder is chosen as the model. Spectrograms are used to convert speech signals into images, serving as input for training the autoencoder. The developed speaker recognition system is compared against traditional systems relying on the MFCC feature extraction technique. In addition to addressing the challenges of a small dataset, the paper explores the impact of a "mismatch condition" by using different time durations of the audio signal during both training and testing phases. Through experiments involving various activation and loss functions, the optimal pair for the small dataset is identified, resulting in a high success rate of 92.4% in matched conditions. Traditionally, Mel-Frequency Cepstral Coefficients (MFCC) have been widely used for this purpose. However, the COVID-19 pandemic has drawn attention to the virus's impact on the human body, particularly on areas relevant to speech, such as the chest, throat, vocal cords, and related regions. COVID-19 symptoms, such as coughing, breathing difficulties, and throat swelling, raise questions about the influence of the virus on MFCC, pitch, jitter, and shimmer features. Therefore, this research aims to investigate and understand the potential effects of COVID-19 on these crucial features, contributing valuable insights to the development of robust speaker recognition systems.