Modified Mel Frequency Cepstral Coefficient Research Articles

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.

Gender of Fetus Identification Using Modified Mel-Frequency Cepstral Coefficients Based on Fractional Discrete Cosine Transform

Gender of Fetus Identification Using Modified Mel-Frequency Cepstral Coefficients Based on Fractional Discrete Cosine Transform

COVID-19: respiratory disease diagnosis with regularized deep convolutional neural network using human respiratory sounds.

Human respiratory sound auscultation (HRSA) parameters have been the real choice for detecting human respiratory diseases in the last few years. It is a challenging task to extract the respiratory sound features from the breath, voice, and cough sounds. The existing methods failed to extract the sound features to diagnose respiratory diseases. We proposed and evaluated a new regularized deep convolutional neural network (RDCNN) architecture to accept COVID-19 sound data and essential sound features. The proposed architecture is trained with the COVID-19 sound data sets and gives a better learning curve than any other state-of-the-art model. We examine the performance of RDCNN with Max-Pooling (Model-1) and without Max-Pooling (Model-2) functions. In this work, we observed that RDCNN model performance with three sound feature extraction methods [Soft-Mel frequency channel, Log-Mel frequency spectrum, and Modified Mel-frequency Cepstral Coefficient (MMFCC) spectrum] for COVID-19 sound data sets (KDD-data, ComParE2021-CCS-CSS-Data, and NeurlPs2021-data). To amplify the models’ performance, we applied the augmentation technique along with regularization. We have also carried out this work to estimate the mutation of SARS-CoV-2 in the five waves using prognostic models (fractal-based). The proposed model achieves state-of-the-art performance on the COVID-19 sound data set to identify COVID-19 disease symptoms. The model’s learnable parameter gradients have vanished in the intermediate layers while optimizing the prediction error which is addressed with our proposed RDCNN model. Our experiments suggested that 3 × 3 kernel size for regularized deep CNN (without max-pooling) shows 2–3% better classification accuracy compared to RDCNN with max-pooling. The experimental results suggest that this new approach may achieve the finest results on respiratory diseases.

A new quantitative acoustic emission model for damage characterization of composite laminates using original waveforms

The correlation between acoustic characteristics and mechanical behaviors shows great significance for health monitoring and characterization. This paper develops a new quantitative model based on the modified Mel-frequency cepstral analysis and statistical methods so as to link acoustic emission (AE) features with mechanical behaviors of end-notched flexure (ENF) composite laminates. First, the Mel-frequency cepstral analysis in automatic speech recognition is modified to adapt to AE sensors and signals. Second, the modified Mel-frequency cepstral coefficients (MFCCs) are extracted from original waveforms of AE hits for damage characterization of composites. MFCC0 is taken as an effective feature to qualitatively discriminate damage stages and to identify the pre-failure critical point. The decreasing patterns of MFCC1 and MFCC2 for ENF specimens can be clearly observed with the loading time by using the simple moving average method. Third, pencil lead breaks are repeatedly conducted on the healthy specimen to verify the pattern in the degraded specimen. Finally, a further investigation based on the cumulative moving average method demonstrates that MFCC1 and MFCC2 are quadratic and linear functions of the load ratio or the deflection ratio, respectively. In addition, the latter is more suitable to be an indicator of damage accumulation of composite laminates.

Motion state classification for micro‐drones via modified Mel frequency cepstral coefficient and hidden Markov mode

In radar detection for micro-drones, the performance of the characteristics and features of micro-drones distinctly varies depending on their motion states. Therefore, this study presents an extraction method for spectral signatures of radar cross section (RCS) to classify and discriminate between typical motion states of micro-drones. The features related to the RCS of the micro-drones are transformed and converted into a modified Mel frequency cepstral coefficient (MFCC), which is proposed to detailedly characterize the spectral change of RCS along with different movements. Then the hidden Markov model (HMM) is used to recognize the typical movements of micro-drones based on the modified MFCC. Experiments in X, Ku bands and HH, VV polarization modes are conducted to investigate the effectiveness of the proposed method, and the classification results with average accuracy above 97 % are also demonstrated in experiments in discriminating between five different motion scenarios of micro-drone. The works involved will aid in maximizing classification accuracy for micro-drones in real scenarios and subsequent situation awareness.

Image-based features for speech signal classification

Like other applications, under the purview of pattern classification, analyzing speech signals is crucial. People often mix different languages while talking which makes this task complicated. This happens mostly in India, since different languages are used from one state to another. Among many, Southern part of India suffers a lot from this situation, where distinguishing their languages is important. In this paper, we propose image-based features for speech signal classification because it is possible to identify different patterns by visualizing their speech patterns. Modified Mel frequency cepstral coefficient (MFCC) features namely MFCC- Statistics Grade (MFCC-SG) were extracted which were visualized by plotting techniques and thereafter fed to a convolutional neural network. In this study, we used the top 4 languages namely Telugu, Tamil, Malayalam, and Kannada. Experiments were performed on more than 900 hours of data collected from YouTube leading to over 150000 images and the highest accuracy of 94.51% was obtained.

Simulation and detection of tamil speech accent using modified mel frequency cepstral coefficient algorithm

Automatic Speech reconstruction system is a topic of interest of many researchers. Since many online courses are come into the picture, so recent researchers are concentrating on speech accent recognition. Many works have been done in this field. In this paper speech accent recognition of Tamil speech from different zones of Tamilnadu is addressed. Hidden Markov Model (HMM) and Viterbi algorithms are very popularly used algorithms. Researchers have worked with Mel Frequency Cepstral Coefficients (MFCC) to identify speech as well as speech accent. In this paper speech accent features are identified by modified MFCC algorithm. The classification of features is done by back propagation algorithm.

Pattern recognition approach to identify loose particle material based on modified MFCC and HMMs

This paper presented a revised particle impact noise detection (PIND) system to detect loose particles in the sealed electronic devices. The type of particle material is the valuable information to trace the particle’ origins and assess the potential harmfulness. The most traditional classification model cannot effectively recognize the types of material due to particle impact acoustic signals with characteristics such as non-stationary, poor repeatability and reproducibility. To address this problem, a new method of pattern recognition for material identification is proposed based on modified Mel frequency cepstral coefficient (MFCC) and hidden Markov model (HMM). First, to be more suitable for target identification of different material of particle acoustic signals, the relationship between the MFCC features and the Mel filter bank is derived, and the modified filter bank is proposed. The modified MFCC features are then mapped into HMMs, where the model number with the maximum probability is chosen as the identified class. Finally, experiments have been carried out with different features and classifiers. The results show that the proposed method can effectively distinguish the four types of materials including wire, chip, aluminum and tin.

A FRAMEWORK FOR MULTILINGUAL TEXT- INDEPENDENT SPEAKER IDENTIFICATION SYSTEM

This article evaluates the performance of Extreme Learning Machine (ELM) and Gaussian Mixture Model (GMM) in the context of text independent Multi lingual speaker identification for recorded and synthesized speeches. The type and number of filters in the filter bank, number of samples in each frame of the speech signal and fusion of model scores play a vital role in speaker identification accuracy and are analyzed in this article. Extreme Learning Machine uses a single hidden layer feed forward neural network for multilingual speaker identification. The individual Gaussian components of GMM best represent speaker-dependent spectral shapes that are effective in speaker identity. Both the modeling techniques make use of Linear Predictive Residual Cepstral Coefficient (LPRCC), Mel Frequency Cepstral Coefficient (MFCC), Modified Mel Frequency Cepstral Coefficient (MMFCC) and Bark Frequency Cepstral Coefficient (BFCC) features to represent the speaker specific attributes of speech signals. Experimental results show that GMM outperforms ELM with speaker identification accuracy of 97.5% with frame size of 256 and frame shift of half of frame size and filter bank size of 40.