Development of language identification system using MFCC and vector quantization

Facial Parts Detection (FPD) approach in conjunction with Vector Quantization (VQ) algorithm are proposed in this paper for face recognition. There are three phases in the proposed system, namely, Preprocessing, Feature Extraction, and Classification. Detecting facial parts, which are nose, both eyes, and mouth, and choosing appropriate dimensions for each part are done in the preprocessing phase. In the feature extraction phase, four groups for each person, one group for each detected part, are constructed for dimensionality reduction and feature discrimination by considering all parts of all training poses. For further data compaction, VQ algorithm employing Kekre Fast Codebook Generation (KFCG) approach for codebook initialization is applied to each of the four groups. Finally, Euclidean distance criterion is used to obtain the recognition rates. Four databases, namely, ORL, YALE, FERET, and FEI that have different facial variations, such as illuminations, rotations, makeups, facial expressions, etc. are used to evaluate the proposed system. Experimental work is performed to evaluate the performance of the proposed technique and the state-of-the-arts approaches. Then, K-Fold Cross Validation (CV) is used to analyze the results. The proposed system consistently improved the recognition rates as well as the storage requirements. Sample results are given.

In speaker recognition, most of the computation originates from the likelihood computations between feature vectors of the unknown speaker and the models in the database. In this paper, we concentrate on optimizing Mel Frequency Cepstral Coefficient (MFCC) for feature extraction and Vector Quantization (VQ) for feature modeling. We reduce the number of feature vectors by pre-quantizing the test sequence prior to matching, and number of speakers by ruling out unlikely speakers during recognition process. The two important parameters, Recognition rate and minimized Average Distance between the samples, depends on the codebook size and the number of cepstral coefficients. We find, that this approach yields significant performance when the changes are made in the number of mfcc's and the codebook size. Recognition rate is found to reach upto 89% and the distortion reduced upto 69%.

This paper addresses the development of an Automatic Speech Recognition (ASR) system for the Moroccan Dialect. Dialectal Arabic (DA) refers to the day-to-day vernaculars spoken in the Arab world. In fact, Moroccan Dialect is very different from the Modern Standard Arabic (MSA) because it is highly influenced by the French Language. It is observed throughout all Arab countries that standard Arabic widely written and used for official speech, news papers, public administration and school but not used in everyday conversation and dialect is widely spoken in everyday life but almost never written. we propose to use the Mel Frequency Cepstral Coefficient (MFCC) features to specify the best speaker identification system. The extracted speech features are quantized to a number of centroids using vector quantization algorithm. These centroids constitute the codebook of that speaker. MFCC’s are calculated in training phase and again in testing phase. Speakers uttered same words once in a training session and once in a testing session later. The Euclidean distance between the MFCC’s of each speaker in training phase to the centroids of individual speaker in testing phase is measured and the speaker is identified according to the minimum Euclidean distance. The code is developed in the MATLAB environment and performs the identification satisfactorily.

This paper presents significance of Mel-frequency Cepstral Coefficients (MFCC) Frequency band selection for text-independent speaker identification. Recent studies have been focused on speaker specific information that may extends beyond telephonic passband. The selection of the frequency band is an important factor to effectively capture the speaker specific information present in the speech signal for speaker recognition. This paper focuses on development of a speaker identification system based on MFCC features which are modeled using vector quantization. Here, the frequency band is varied up to 7.75 kHz. Speaker identification experiments evaluated on TIMIT database consisting of 630 speaker shows that the average recognition rate achieved is 97.37 % in frequency band 0–4.85 kHz for 20 MFCC filters.

Objective To investigate the feasibility of speech recognition technologies including double threshold method, Mel frequency cepstral coefficient (MFCC) and vector quantization (VQ) in automatic recognition of cough sounds. Methods Five healthy adults and 15 patients with cough were recruited for recording of both non- cough and cough sounds in a quiet environment. The records were randomized into training and testing samples. The training samples were used to generate the code book for cough recognition software, which was then used to recognize and analyze the testing samples automatically.The sensitivity and specificity of recognition were calculated by comparison with outcomes from human ear recognition. The recognition time in two approaches was recorded. Results Two hundred cough and 200 noncough sound samples were used to generate the code book for cough recognition software, while 375 cough sound samples and 125 non-cough sound samples were used as the testing samples. The recognition time of the testing samples needed was 33 minutes and 18 seconds by human ear recognition vs 1 minute and 35seconds by code book-based automatic recognition. In addition, the sensitivity and specificity in code bookbased automatic recognition of the cough sound were 98.93% and 100% respectively. Conclusion The double threshold method based on VQ and MFCC appears feasible in automatic recognition of cough sounds. Key words: Cough; Speech recognition software; Pattern recognition, automatic; Sensitivity and specificity; Human ear recognition

Competitive learning algorithms for vector quantization

This paper presents the implementation of a practical voice recognition system using MATLAB (R2014b) to secure a given user’s system so that only the user may access it. Voice recognition systems have two phases, training and testing. During the training phase, the characteristic features of the speaker are extracted from the speech signal and stored in a database. In the testing phase, the stored audio features of the test voice sample are compared with the voice samples in the database and determined if a match exists. For this research, Mel Frequency Cepstral Coefficients (MFCCs) were chosen to represent the feature vectors of the user’s voice as it accurately simulates the behavior of the human ear. This characteristic of the MFCCs makes them an excellent measure of speaker characteristics. The feature matching process is then performed by subjecting the MFCCs to vector quantization using the LBG (Linde-Buzo-Gray) algorithm. In practical scenarios, noise is a major factor that adversely influences a voice recognition system. The paper addresses this issue by utilizing spectral subtraction to remove environmental noise affecting the speech signal thereby increasing the robustness of the system.

The goal of tone recognition is to accurately identify the type and the name of the musical instruments through processing and analyzing the sound signals. In order to reduce the influence of feature confusion on classification process, a method of tone recognition based on the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) and Random Forests is proposed. In this process, Mel Frequency Cepstral Coefficients (MFCCs) are acquired, and the quadratic sum of distance between two MFCCs and the entropy of information are computed which are used as indices to analyze and select the MFCCs based the TOPSIS decision. The selected MFCCs are used to classify tone of trumpet- piano, trumpet-cello and piano-cello, and the recognition rates were 100%, 99.9% and 100% respectively. The results are satisfactory and verify feasibility of the developed method.

In order to improve the speaker recognition accuracy, a new mel-frequency cepstral coefficient (MFCC) parameters extraction criterion based on F-Ratio and correlated distance criterion is proposed. Take the correlation of the characteristics into account, F-Ratio and correlated distance criterion (FRCDC) is proposed, which provides an idea about the significance of each component for MFCC feature. Two methods based on this criterion to extract MFCC parameters are given: dimensionality reduced according to the FRCDC value, improving the MFCC's discrimination by using a new bandpass liftering window. Experimental results based on two languages voice database show that the recognition rate adopting proposed dimensionality reduced method is improved by 10-15% than that of method without dimensionality reduced. The recognition rate using the new liftering process is improved by 20-25% than that without the liftering and 10-20% than conventional methods.

Speaker recognition is a process of recognizing a person based on their unique voice signals and it is a topic of great importance in areas of intelligent and security. Considerable research and development has been carried out to extract speaker specific features and to develop features matching techniques. The goal of this paper is to perform text-independent speaker identification. These models rely on Mel Frequency Cepstral Coefficients (MFCC) for extraction of speaker specific features and for speaker modelling Vector Quantization (VQ) is used due to high accuracy and simplicity. The proposed system efficiency was analyzed by using 20 filter banks for extracting features. The performance was evaluated using MATLAB against different speakers in different languages such as Tamil, Malayalam, Hindi, Telugu and English with duration of 2, 3 and 4 s. Experimental result shows that 4 s duration of speech regardless of language is able to produce 98 %, 99 % and 97 % of identification when compared to 2 and 3 s. The system efficiency may further be improved using other speaker modelling techniques like Neural Network, Hidden Markov Model and Gaussian Mixture Model.KeywordsSpeaker IdentificationMel-frequency Cepstral Coefficients (MFCC)Speaker RecognitionSpeaker Modeling TechniquesAutomatic Speaker Recognition SystemThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

This paper introduces a new approach to control and drive the DC motor, using speech recognition. The speech signal can be provided through microphone that is connected to computer. A DC motor connected through microcontroller can be driven in forward or reverse direction at different speeds, as well as it can be stopped by giving speech command. In this paper Mel Frequency Cepstral Co-efficient (MFCC) is used to recognize the user's speech and vector quantization is used to increase speech recognition accuracy. The MFCC and VQ algorithm, for speech recognition have been implemented in MATLAB 7.5 version on Windows Vista platform, and the supporting hardware setup is being implemented.

In this paper, we propose a temporal dynamic index reordering vector quantization for wavelet-based embedded coding. Da Silva et al. introduced a vector quantization concept that is similar to EZW called a successive approximation vector quantizer (SAVQ). The successive refinement process is defined as a temporal process in our proposed algorithm. The temporal updates are performed in every refinement pass, and the updates of codevectors reflect the updates of angles for vector approximation. Because the approximation trajectory of SAVQ is similar to that of the least-mean- square algorithm, temporal redundancy does not seem to be obvious. However, the redundancy becomes more clear in the improved SAVQ. By carefully analyzing the angle transitions, we are successfully able to apply dynamic index reordering vector quantization (DIRVQ) in the temporal domain and improve the coding performance. Efficient encoding and decoding algorithms for D4 lattice vector quantization are also proposed, and the algorithms can be applicable to DIRVQ to reduce the computational complexity of the reordering process. Experiments are performed with the 2x2 vector size. For almost all tested bpp's the PSNR performance of our proposed algorithm for a lena image outperforms that of SAVQ up to 0.18 dB. Improvement is measured for the highly detailed baboon image as well.

This paper will describe the vector quantization (VQ) of speech spectrum parameters as used in state of the art real time speech coding algorithms. The complexity of these speech coding algorithms, combined with the constraints of real time operation, limit the practical size of the vector quantization tables. This conflicts with the requirement that large VQ tables are needed to provide high quality coded speech. The goal of this paper is to demonstrate the feasibility of using tree searched vector quantizers to overcome this conflict.

Recognition of car horns based on principal component analysis of MEL frequency Cepstral coefficients and support vector machine

In the study of speaker identification, Mel Frequency Cepstral Coefficient (MFCC) method is the best and most popular which is used to feature extraction. Further Vector Quantization (VQ) technique is used to minimize the amount of data to be handled and mapping vectors from a large vector space to a finite number of regions in that space in recent years. Voice, like other biometrics, cannot be forgotten or misplaced, unlike knowledge-based (e.g., password) or possession-based (e.g., key) access control methods. In the present work, modified Mel frequency cepstral coefficients using discrete fractional Fourier transform and vector quantization is obtained. The experimental results are analyzed with the help of MATLAB.