Hidden Markov Model Toolkit Research Articles

An automatic speech recognition system in Indian and foreign languages: A state-of-the-art review analysis

Speech Recognition is one of the prominent research topics in the field of Natural Language Processing (NLP). The Speech Recognition technique removes the barriers and makes the system ease for inter-communication between human beings and devices. The aim of this study is to analyze the Automatic Speech Recognition System (ASRS) proposed by different researchers using Machine learning and Deep Learning techniques. In this work, Indian and foreign languages speech recognition systems like Hindi, Marathi, Malayalam, Urdu, Sanskrit, Nepali, Kannada, Chinese, Japanese, Arabic, Italian, Turkish, French, and German are considered. An integrated framework is presented and elaborated with recent advancement. The various platform like Hidden Markov Model Toolkit (HMM Toolkit), CMU Sphinx, Kaldi toolkit are explained which is used for building the speech recognition model. Further, some applications are elaborated which depict the uses of ASRS.

A Hidden Markov Model-Based Speech Recognition System Using Baum-Welch, Forward-Backward and Viterbi Algorithms

Speech is the most complex part or component of human intelligence and for that matter speech signal processing is very important. The variability of speech is very high, and this makes speech recognition difficult. Other factors like dialects, speech duration, context dependency, different speech speed, speaker differentiation, environment and locality all add to the difficulty in speech processing. The absence of distinct boundaries between tones or words causes additional problems. Speech has speaker dependent characteristics, so that no one can reproduce or repeat phrases in the same way as another. Nevertheless, a speech recognition system should be able to model and recognize the same words and phrases absolutely. Digital signal processors (DSP) are often used in speech signal processing systems to control these complexities. This paper presents a Hidden Markov Model (HMM) based speech signal modulation through the application of the Baum-Welch, Forward-Backward and Viterbi algorithms. The system was implemented using a 16-bit floating point DSP (TMS320C6701) from Texas instruments and the vocabulary was trained using the Microsoft Hidden Markov Model Toolkit (HTK). The proposed system achieved about 79% correct word recognition which represents approximately 11,804 correct words recognized out of a total of 14960 words provided. This result indicates that the proposed model accuracy and speaker independent system has a very good evaluation score, and thus can be used to aid dictation for speech impaired persons and applications in real time with a 10 ms data exchange rate.

Histogram of Marked Background (HMB) Feature Extraction Method for Arabic Handwriting Recognition

Features extraction is one of the most important steps in handwriting recognition systems. In this paper, we propose a novel features extraction method, which is adapted to the complex nature of Arabic handwriting. The proposed feature called histogram of marked background (HMB) is not considering only ink pixels in a text image, but also uses the background of the image. Each background pixel in the text image was marked according to the repartition of ink pixels in its neighborhood. Feature vectors are extracted by computing histograms from the marked images. Hidden Markov models (HMMs) with Hidden Markov model toolkit (HTK) were used in the recognition process. The experiments were performed on two datasets: IBN SINA database of historical Arabic documents and Isolated Farsi Handwritten Character Database (IFHCDB). The proposed feature in this study produced efficient and promising results for Arabic handwriting recognition, for both isolated characters and for historical documents.

Optimal parameters selected for automatic recognition of spoken Amazigh digits and letters using Hidden Markov Model Toolkit

In this paper, we present our Amazigh automatic speech recognition system. Its realization is constructed with context-independent phonetic Hidden Markov Models. Many choices are made on this system, such as the number of states of the models, the type of emission probability densities associated with the states, and the representation of the signal by cepstral coefficients. The results of recognition of our system place it at a level of height performance comparable to that achieved by Markovian automatic speech recognition systems. Our system is designed to recognize 43 distinct isolated Amazigh words (33 letters and 10 digits). The recognition rate is then calculated for each digit and letter. The overall accuracy and word recognition rate for the whole database achieved 91.31% after extensive testing and change of the recognition parameters. The results obtained in this work are improved in association with our previous work concerning Amazigh spoken digits and letters automatic speech recognition, using Hidden Markov Model Toolkit.

Automatic Speech Recognition (ASR) System for Isolated Marathi Words: Using HTK

The present manuscript focuses on building automatic speech recognition (ASR) system for Marathi language (M-ASR) using Hidden Markov Model Toolkit (HTK). The M-ASR system gives the detail about experimentation and implementation using the HTK Toolkit. In this work total 106 speaker independent Marathi isolated words were recognized. These unique Marathi words are used to train and evaluate M-ASR system. The speech corpus (database) is created by us using isolated Marathi words uttered with mixed gender people. The system uses Mel Frequency Cepstral Coefficient (MFCC) for the purpose of extracting features using Gaussian mixture model (GMM). Viterbi algorithm based on token passing is used for decoding to recognize unknown utterances. The proposed M-ASR system is speaker independent. The proposed system has reported 96.23% word level recognition accuracy.

SEGMENTATION-FREE RECOGNITION OF URDU SCRIPT USING HMM

All the Urdu literature is in the form of manuscripts and typewritten books.There is a need for converting all these physical libraries into electronic libraries. Various OCRs have been developed for different languages and are widely used. Building a complete Urdu OCR is a difficult task because Urdu is highly cursive language, where ligatures overlap and style variation poses challenges to the recognition system. We are describing a technique for automatic recognition of off-line printed Urdu text using Hidden Markov Models. Our method does not require segmentation into characters and considers each shape of Urdu character as different class resulting in a total of 196 classes (compared to 38 Urdu letters). This paper presents a novel feature extraction method based on sliding window technique, using only 16 statistical features from each sliding window thereby eliminating the need for segmentation of Urdu text. The dependency of Recognition rate of Urdu script upon, the number of states of HMM, different sizes of hierarchical window and different fonts is presented. We are using HTK (Hidden Markov Model Toolkit) for training, recognition and result analysis.

A Novel Technique for Speech Recognition and Visualization Based Mobile Application to Support Two‐Way Communication between Deaf‐Mute and Normal Peoples

Mobile technology is very fast growing and incredible, yet there are not much technology development and improvement for Deaf‐mute peoples. Existing mobile applications use sign language as the only option for communication with them. Before our article, no such application (app) that uses the disrupted speech of Deaf‐mutes for the purpose of social connectivity exists in the mobile market. The proposed application, named as vocalizer to mute (V2M), uses automatic speech recognition (ASR) methodology to recognize the speech of Deaf‐mute and convert it into a recognizable form of speech for a normal person. In this work mel frequency cepstral coefficients (MFCC) based features are extracted for each training and testing sample of Deaf‐mute speech. The hidden Markov model toolkit (HTK) is used for the process of speech recognition. The application is also integrated with a 3D avatar for providing visualization support. The avatar is responsible for performing the sign language on behalf of a person with no awareness of Deaf‐mute culture. The prototype application was piloted in social welfare institute for Deaf‐mute children. Participants were 15 children aged between 7 and 13 years. The experimental results show the accuracy of the proposed application as 97.9%. The quantitative and qualitative analysis of results also revealed that face‐to‐face socialization of Deaf‐mute is improved by the intervention of mobile technology. The participants also suggested that the proposed mobile application can act as a voice for them and they can socialize with friends and family by using this app.

Interface of an Automatic Recognition System for Dysarthric Speech

This paper addresses the realization of a Human/Machine (H/M) interface including a system for automatic recognition of the Continuous Pathological Speech (ARSCPS) and several communication tools in order to help frail people with speech problems (Dysarthric speech) to access services providing by new technologies of information and communication (TIC) while making it easier for the doctors to achieve a first diagnosis on the patient’s disease. In addition, an ARSCPS has been improved and developed for normal and pathology voice while establishing a link with our graphic interface which is based on the box tools Hidden Markov Model Toolkit (HTK), in addition to the Hidden Models of Markov (HMM). In our work we used different techniques of feature extraction for the speech recognition system in order to improve the dysarthric speech intelligibility while developing an ARSCPS which can perform well for pathological and normal speakers. These techniques are based on the coefficients of ETSI standard Mel Frequency Cepstral Coefficient Front End (ETSI MFCC FE V2.0); Perceptual Linear Prediction coefficients (PLP); Mel Frequency Cepstral Coefficients (MFCC) and the recently proposed Power Normalized Cepstral Coefficients (PNCC) have been used as a basis for comparison. In this context we used the Nemours database which contains 11 speakers that represents dysarthric speech and 11 speakers that represents normal speech.

Development and Analysis of Speech Recognition Systems for Assamese Language Using HTK

Language analysis is very important for the native speaker to connect with the digital world. Assamese is a relatively unexplored language. In this report, we analyze different aspects of speech-to-text processing, starting from building a speech corpus, defining syllable rules, and finally developing a speech search engine of Assamese. We have collected about 20 hours of speech in three (viz., read, extempore, and conversation) modes and transcribed it. We also discuss some issues and challenges faced during development of the corpus. We have developed an automatic syllabification model with 11 rules for the Assamese language and found an accuracy of more than 95% in our result. We found 12 different syllable patterns where 5 are found most frequent. The maximum length of a syllable found is four letters. With the help of Hidden Markov Model Toolkit (HTK) 3.5, we used deep learning based neural network for our speech recognition model, where we obtained 78.05% accuracy for automatic transcription of Assamese speech.

Arabic Speech Recognition for Connected Words Using HTK: Triphones Expanded to Gmm Based Quran Recognition

In this paper, the authors describe and propose an efficient and effective method of speaker-independent continuous Arabic speech recognition method, based on a phonetically rich speech corpus. This speech corpus contains two datasets: (1) the Spoken Arabic Digits (SAD), recorded by 66 speakers (33 male and 33 female), holds 6600 digit words in which each speaker pronounces each digit from zero to nine ten times, and (2) the second dataset is restrained to quranic sentences of three famous speakers with Tajweed rules (somehow of singing) of the last 30 chapters (Surat) of the Holy Quran. The problematic of co-articulation was treated in general with triphones as acoustic model and bigram as language model which was the most appropriate in this text; and the adjustment’s problem of sound duration in particular was also treated, with triphones expanded to Gaussians mixtures models (GMM). The proposed Arabic speech recognition system is based on the Cambridge Hidden Markov Model (HMM) Toolkit (HTK) tools. The Experimental tests show that the digit dataset using 3 emitting states per phone, has an excellent word recognition (WR) rate of 97.95% and sentence recognition (SR) rate of 93.14%. The best result of Quranic phrases obtained is 73.44% of WR rate and 14.38% of SR rate with 66.09% of accuracy within “Elmirigli” reader. Adapting the basic system to the speaker’s intonation voice, then applying an appropriate GMM to each tied-list triphone, it outperforms the Hidden Markov Models HMM-triphone acoustic model experiment by 16.93% of WR rate within “Alsudaissi” reader thanks to the solved duration vowels (mudud) problem. The contribution of this work is to apply triphones expanded to Gmms method which was compared to Maximum Likelihood Linear Regression (MLLR) and beat it by almost 5% in “Alsudaissi” reader.

Development of Standard Yorùbá speech-to-text system using HTK

In this paper, a Standard Yoruba speech-to-text system capable of recognizing isolated words spoken by the users based on previously stored data was designed and implemented. This system adopted syllable-based approach, and carefully-selected words were recorded, analyzed and annotated, using Praat software. An experimental database of six native speakers was taken, each speaking 25 bi-syllabic and 25 tri-syllabic words, under an acoustically-controlled room. The meaningful spectral coefficients were successfully extracted using Mel-frequency cepstral coefficients technique and Hidden Markov Model Toolkit was used to implement the system. A graphical user interface was also developed to make the system accessible and more interactive. Furthermore, the system was tested and evaluated based on the perception of native speakers of the language. The overall accuracy for bi-syllabic and tri-syllabic words was 76 and 84 % respectively. These results obtained for both bi and tri-syllabic words showed that this system was a promising approach that could be adopted for Standard Yoruba continuous speech recognition system as this will make the system useable for the foreign speaker.

Using features of local densities, statistics and HMM toolkit (HTK) for offline Arabic handwriting text recognition

This paper presents an analytical approach of an offline handwritten Arabic text recognition system. It is based on the Hidden Markov Models (HMM) Toolkit (HTK) without explicit segmentation. The first phase is preprocessing, where the data is introduced in the system after quality enhancements. Then, a set of characteristics (features of local densities and features statistics) are extracted by using the technique of sliding windows. Subsequently, the resulting feature vectors are injected to the Hidden Markov Model Toolkit (HTK). The simple database “Arabic-Numbers” and IFN/ENIT are used to evaluate the performance of this system.

Automatic Speech Segmentation Using Hybrid Wavelet Features and HMM

In this research, a novel feature set is used to automatically segment speech signal. Automatic segmentation is veryuseful especially for large database. A hybrid features model is created from wavelet packet analysis and mel-scale is used to train Hidden Markov Model (HMM) for phone boundary detection. HMM is implemented using the Hidden Markov Model Toolkit (HTK).The database (Ked-TIMIT) is used for result verifications and Mel Frequency Cepstral Coefficients (MFCC) is used as reference for evaluating the results of the proposed Hybrid model. The results are categorized for vowels, consonants and short phones. Phone duration and start location are used as metrics to evaluate the system success rate. Success rate of 74% is achieved for consonant detection, 72% for vowel detection and 58% for short phone detection. Using the simple metric that relies only on boundary locations but ignoring duration, the achieved results are 92.5% for consonant detection, 90% for vowel detection and 77.5% for short phoneme detection. In addition to boundary detection the proposed hybrid model is utilized to compare newly developed features called Mel scale Best Tree Encoding (Mel-BTE ) to the mostly used popular features MFCC along with all experiments using the same database. The relative results for Mel-BTE with respect to MFCC are 94.77% for consonant detection, 87.5% for vowel detection and 93.33% for short phoneme detection.

Isolated Digits Recognition in Kannada Language

In this paper, have implemented the isolated digit recognition in Kannada language using Hidden Markov Model Toolkit (HTK). Hidden Markov models used as pattern recognizer with the help of MFCC as a featured vector of the wave samples. The paper focuses on all isolated digits of Kannada i.e. Sonne to Ommbattu (0 to 9), The system helps in interaction of rural people and the computer or any system. The system data structure is defined at word level and its performance is evaluated.

Feature Extraction Method for Improving Speech Recognition in Noisy Environments

The paper presents a feature extraction method, named as Normalized Gammachirp Cepstral Coefficients (NGCC) that incorporates the properties of the peripheral auditory system to improve robustness in noisy speech recognition. The proposed method is based on a second order low-pass filter and normalized gammachirp filterbank to emulate the mechanisms performed in the outer/middle ear and cochlea. The speech recognition performance of this method is conducted on the speech signals in real-world noisy environments. Experimental results demonstrate that method outperformed the classical feature extraction methods in terms of speech recognition rate. The used Hidden Markov Models based speech recognition system is employed on the HTK 3.4.1 platform (Hidden Markov Model Toolkit).

Extraction of Speech Parameters from Speech Database using Festival

Speech synthesis is the process of production of artificial speech. The system used for generation of speech from text is called as text-to-speech (TTS) system. In TTS system, text and voice models for a particular language or multiple languages are given as input to the system, which generates speech as output corresponding to the provided voice models. Speech synthesis systems can be extremely useful to people who are visually challenged, visually impaired and illiterate to get into the mainstream society. More recent applications include spoken dialogue systems and communicative robots. HMM (Hidden Markov Model) based Speech synthesis is the emerging technology for TTS. HMM based speech synthesis system consists of training phase and synthesis phase. In the training part, phone and excitation parameters are extracted from speech database and modeled by context dependent HMMs. In synthesis part, the system will extract the suitable phone and excitation parameters from the previously trained models and generates the speech. The main objective of this project is to build an HMM based speech synthesis system. In the training process, the system uses HTK (Hidden Markov Model Tool Kit) and SPTK (Signal Processing Tool Kit) developed at Cambridge University and Tokyo Institute of Technology respectively. Synthesis part is be done by „Festival‟. Festival is a speech synthesis tool for the generation of speech and it is language independent which is developed at the University of Edinburgh. The main advantage of this approach is its flexibility in changing speaker identities, emotions and speaking styles.

Analysis of a Small Vocabulary Bangla Speech Database for Recognition

To carry out any kind of research in the field of speech signal processing, a standard database is essential. There are many databases in different languages but not in Bangla language. Therefore, in this article, it has been attempted to develop and analysis a small vocabulary Bangla database for recognition. In this database 11 Bangla digits (/ak/, /dui/, /tin/, /chaar/, /panch/, /chhoy/, /shaat/, /aat/, /noy/, /zero/, /shunno/) have been used. The developed database consisted of two sets of data such as training and testing datasets. The training dataset contains 3824 utterances of 50 speakers, and testing dataset is subdivided into four groups (clean1, clean2, clean3 and clean4) and contains 1985 utterances of 52 speakers. All recordings have been done in a quiet room but not sound proof with the A4Tech HS-60 headset microphone interfaced to an Intel Dual Core 2.0 GHz CPU. The software used to record and edit the speech file is wavepad. Finally, an HMM based recognizer is developed to evaluate the database. The word accuracy for test sets is found to be 98.05% on the average. In this recognition process Mel-LPC based front-end and as a reference recognizer HTK (Hidden Markov Model Toolkit) have been used. General Terms Bangla Speech Database Processing and Recognition.

Automatic Speech Recognition of Pathological Voice

Background/Objectives: Automatic speech recognition (ASR) benefits human beings in many useful applications. Various ASR systems exhibiting good performance have been developed for normal speakers. The speech produced by a voice disordered patient is not like a normal speaker due to irregular vibration and incomplete closure of vocal fold. Therefore, an investigation is required by exploring the different speech features to develop an ASR system which can perform well for both pathological and normal speakers. Methods: In this paper, we proposed an automatic speech recognition system using Hidden Markov Model Toolkit (HTK) for normal and pathology voice. Four techniques are applied for feature extraction; Mel Frequency Cepstral Coefficient (MFCC), Perceptual Linear Prediction (PLP), RelAtiveSpecTrA - Perceptual Linear Predictive (RASTA-PLP), and linear prediction coefficients (LPC). The database that used to evaluate the performance of the developed system; includes a total of 297 speakers 121 of them were normal speakers and the remaining containing five types of vocal fold disorders. Findings: Experimental results show that the developed system gives good accuracies for normal and pathology voice. The highest accuracy of 94.44 % with a word error rate 5.55% is achieved in case of normal voice, and 88.63 % with a word error rate 11.63 % in case of pathology voice. Fuzzy logic controller is proposed to automatically segmentation the normal and disorders voice.

Comparative Analysis of the Vocal Repertoire of Eulemur: A Dynamic Time Warping Approach

The diversity of qualitative approaches and analytical methods has often undermined comparative research on primate vocal repertoires. The purpose of the present work is to introduce a quantitative method based on dynamic time warping to the study of repertoire size in Eulemur spp. We obtained a large sample of calls of E. coronatus, E. flavifrons, E. fulvus, E. macaco, E. mongoz, E. rubriventer, and E. rufus, recorded between 1999 and 2013 from captive and wild lemurs. We inspected recordings visually using spectrograms, then cut and saved high-quality vocal emissions to single files for further analysis. We extracted the acoustic features of all vocalizations of a species using the Hidden Markov Model Toolkit, an application of dynamic time warping, and then compared cepstral coefficients (a feature widely used in automatic speaker recognition) pairwise. We analyzed the results using affinity propagation clustering. We found that Eulemur species share most of their vocal repertoire but species-specific calls determine repertoire size differences. Repertoire size varied from 9 to 14 vocalization types among species, with a mean of 11.

Automatic detection of the expiratory and inspiratory phases in newborn cry signals

An analysis of newborn cry signals, either for the early diagnosis of neonatal health problems or to determine the category of a cry (e.g., pain, discomfort, birth cry, and fear), requires a primary and preliminary preprocessing step to quantify the important expiratory and inspiratory parts of the audio recordings of newborn cries. Data typically contain clean cries interspersed with sections of other sounds (generally, the sounds of speech, noise, or medical equipment) or silence. The purpose of signal segmentation is to differentiate the important acoustic parts of the cry recordings from the unimportant acoustic activities that compose the audio signals. This paper reports on our research to establish an automatic segmentation system for newborn cry recordings based on Hidden Markov Models using the HTK (Hidden Markov Model Toolkit). The system presented in this report is able to detect the two basic constituents of a cry, which are the audible expiratory and inspiratory parts, using a two-stage recognition architecture. The system is trained and tested on a real database collected from normal and pathological newborns. The experimental results indicate that the system yields accuracies of up to 83.79%.