Speech Recognition Method Research Articles

A Rule-Based Grapheme-to-Phoneme Conversion System

This article presents a rule-based grapheme-to-phoneme conversion method and algorithm for Polish. It should be noted that the fundamental grapheme-to-phoneme conversion rules have been developed by Maria Steffen-Batóg and presented in her set of monographs dedicated to the automatic grapheme-to-phoneme conversion of texts in Polish. The author used previously developed rules and independently developed the grapheme-to-phoneme conversion algorithm.The algorithm has been implemented as a software application called TransFon, which allows the user to convert any text in Polish orthography to corresponding strings of phonemes, in phonemic transcription. Using TransFon, a phonemic Polish language corpus was created out of an orthographic corpus. The phonemic language corpusallows statistical analysis of the Polish language, as well as the development of phoneme- and word-based language models for automatic speech recognition using statistical methods. The developed phonemic language corpus opens up further opportunities for research to improve automatic speech recognition in Polish. The development of statistical methods for speech recognition and language modelling requires access to large language corpora, including phonemic corpora. The method presented here enables the creation of such corpora.

Emotional Speech Recognition Method Based on Word Transcription

The emotional speech recognition method presented in this article was applied to recognize the emotions of students during online exams in distance learning due to COVID-19. The purpose of this method is to recognize emotions in spoken speech through the knowledge base of emotionally charged words, which are stored as a code book. The method analyzes human speech for the presence of emotions. To assess the quality of the method, an experiment was conducted for 420 audio recordings. The accuracy of the proposed method is 79.7% for the Kazakh language. The method can be used for different languages and consists of the following tasks: capturing a signal, detecting speech in it, recognizing speech words in a simplified transcription, determining word boundaries, comparing a simplified transcription with a code book, and constructing a hypothesis about the degree of speech emotionality. In case of the presence of emotions, there occurs complete recognition of words and definitions of emotions in speech. The advantage of this method is the possibility of its widespread use since it is not demanding on computational resources. The described method can be applied when there is a need to recognize positive and negative emotions in a crowd, in public transport, schools, universities, etc. The experiment carried out has shown the effectiveness of this method. The results obtained will make it possible in the future to develop devices that begin to record and recognize a speech signal, for example, in the case of detecting negative emotions in sounding speech and, if necessary, transmitting a message about potential threats or riots.

Research on the Construction of Human-Computer Interaction System Based on a Machine Learning Algorithm

In this paper, we use machine learning algorithms to conduct in-depth research and analysis on the construction of human-computer interaction systems and propose a simple and effective method for extracting salient features based on contextual information. The method can retain the dynamic and static information of gestures intact, which results in a richer and more robust feature representation. Secondly, this paper proposes a dynamic planning algorithm based on feature matching, which uses the consistency and accuracy of feature matching to measure the similarity of two frames and then uses a dynamic planning algorithm to find the optimal matching distance between two gesture sequences. The algorithm ensures the continuity and accuracy of the gesture description and makes full use of the spatiotemporal location information of the features. The features and limitations of common motion target detection methods in motion gesture detection and common machine learning tracking methods in gesture tracking are first analyzed, and then, the kernel correlation filter method is improved by designing a confidence model and introducing a scale filter, and finally, comparison experiments are conducted on a self-built gesture dataset to verify the effectiveness of the improved method. During the training and validation of the model by the corpus, the complementary feature extraction methods are ablated and learned, and the corresponding results obtained are compared with the three baseline methods. But due to this feature, GMMs are not suitable when users want to model the time structure. It has been widely used in classification tasks. By using the kernel function, the support vector machine can transform the original input set into a high-dimensional feature space. After experiments, the speech emotion recognition method proposed in this paper outperforms the baseline methods, proving the effectiveness of complementary feature extraction and the superiority of the deep learning model. The speech is used as the input of the system, and the emotion recognition is performed on the input speech, and the corresponding emotion obtained is successfully applied to the human-computer dialogue system in combination with the online speech recognition method, which proves that the speech emotion recognition applied to the human-computer dialogue system has application research value.

The Concurrent OLSA Test: A Method for Speech Recognition in Multi-talker Situations at Fixed SNR.

A multi-talker paradigm is introduced that uses different attentional processes to adjust speech-recognition scores with the goal of conducting measurements at high signal-to-noise ratios (SNR). The basic idea is to simulate a group conversation with three talkers. Talkers alternately speak sentences of the German matrix test OLSA. Each time a sentence begins with the name “Kerstin” (call sign), the participant is addressed and instructed to repeat the last words of all sentences from that talker, until another talker begins a sentence with “Kerstin”. The alternation of the talkers is implemented with an adjustable overlap time that causes an overlap between the call sign “Kerstin” and the target words to be repeated. Thus, the two tasks of detecting “Kerstin” and repeating target words are to be done at the same time. The paradigm was tested with 22 young normal-hearing participants (YNH) for three overlap times (0.6 s, 0.8 s, 1.0 s). Results for these overlap times show significant differences, with median target word recognition scores of 88%, 82%, and 77%, respectively (including call-sign and dual-task effects). A comparison of the dual task with the corresponding single tasks suggests that the observed effects reflect an increased cognitive load.

ОСОБЛИВОСТІ ВИКОРИСТАННЯ MEL-КЕПСТРАЛЬНИХ КОЕФІЦІЄНТІВ ПРИ РОЗПІЗНАВАННІ МОВЛЕННЯ

One of the main forms of human interaction is speech. Nowadays, there are enough useful programs for speech recognition of people who have certain limitations. These programs translate the text that was spoken aloud into the text, to clearly understand what the person wants to say. The main purpose of speech recognition methods is to obtain information as an input voice signal for further clear translation. Modern methods of speech recognition use an important part - language modeling and acoustic modeling. Nowadays, there are quite a number of methods for translating a voice into a test. The best method of recognition is bas on hidden Markov models. Statistical models are using to display sequences of characters. Therefore, the method based on hidden Markov models is one of the best for solving such problems. The article investigates the methods of speech recognition of people with speech disorders in a short dictionary using mel-keptral coefficients. The proposed method applies the criterion used for unbiased estimation of the logarithmic spectrum to the spectral model represented by MEL coefficients. To solve the nonlinear minimization problem involved in the method, they provide an iterative algorithm whose convergence is guaranteed. Examples of speech analysis and results of an isolated word recognition experiment.

О методе распознавания голосовых команд с применением особого преобразования спектральных плотностей

This article discusses new approach to solving specific problem of speech recognition. The problem is formulated as problem of command recognition. Commands have equal lengths (in words). Each word position has its own set of word candidates. There were developed the models for solving this problem. The models consist of spectral density estimator, convolutional neural networks and naive Bayes classifier. The author performed a comparative analysis of the models and selected the best. There was developed a graphical user interface for interacting with the best model. The result can be used as a base for creation speech recognition methods for voice user interfaces.

Analytical Survey : Speech Recognition Methods Used In Voice Recognition Techniques


 
 
 
 This paper mainly studies Speech recognition methods used in voice control systems Analytical survey, Construction of the system voice control is now urgent. Such systems can greatly facilitate user interaction with the computer system.This is especially the idea developed in the concept of Moreover, sometimes the voice Interface is a necessary component of,for example, when it comes to people with disorders of the musculoskeletal system., The work proposed describes general methods of speech recognition based on hidden Markova models (HMM), for in ability to design control system, has high efficiency through voice commands save time and effort and shorten the distance through this circuit get a security system depends on it registered only the person who by scoring and Help people with special needs to rely on themselves in some work
 
 © 2018 JASET, International Scholars and Researchers Association
 
 
 

Thru-the-wall Eavesdropping on Loudspeakers via RFID by Capturing Sub-mm Level Vibration

The unprecedented success of speech recognition methods has stimulated the wide usage of intelligent audio systems, which provides new attack opportunities for stealing the user privacy through eavesdropping on the loudspeakers. Effective eavesdropping methods employ a high-speed camera, relying on LOS to measure object vibrations, or utilize WiFi MIMO antenna array, requiring to eavesdrop in quiet environments. In this paper, we explore the possibility of eavesdropping on the loudspeaker based on COTS RFID tags, which are prevalently deployed in many corners of our daily lives. We propose Tag-Bug that focuses on the human voice with complex frequency bands and performs the thru-the-wall eavesdropping on the loudspeaker by capturing sub-mm level vibration. Tag-Bug extracts sound characteristics through two means: (1) Vibration effect, where a tag directly vibrates caused by sounds; (2) Reflection effect, where a tag does not vibrate but senses the reflection signals from nearby vibrating objects. To amplify the influence of vibration signals, we design a new signal feature referred as Modulated Signal Difference (MSD) to reconstruct the sound from RF-signals. To improve the quality of the reconstructed sound for human voice recognition, we apply a Conditional Generative Adversarial Network (CGAN) to recover the full-frequency band from the partial-frequency band of the reconstructed sound. Extensive experiments on the USRP platform show that Tag-Bug can successfully capture the monotone sound when the loudness is larger than 60dB. Tag-Bug can efficiently recognize the numbers of human voice with 95.3%, 85.3% and 87.5% precision in the free-space eavesdropping, thru-the-brick-wall eavesdropping and thru-the-insulating-glass eavesdropping, respectively. Tag-Bug can also accurately recognize the letters with 87% precision in the free-space eavesdropping.

Survey of Speaker Adaptation Methods in Speech Recognition

Survey of Speaker Adaptation Methods in Speech Recognition

Machine learning of transcripts and audio recordings of spontaneous speech for diagnosis of Alzheimer’s disease

AbstractBackgroundLanguage deficits may occur in early stages of Alzheimer’s disease (AD). Analysis of individuals’ language changes has been shown to predict the diagnosis and severity of AD. This work explored machine learning methods to detect AD using voice recordings of the Cookie Theft picture description tasks.MethodAn open‐source dataset from Dementia Bank, which included manual transcripts and audio recordings of the Cookie Theft picture description task from 309 individuals with AD and 243 cognitively healthy older adults, was employed in the classification analysis. We analyzed transcripts generated using three different approaches (Figure 1), including a) automatically generated transcripts based on denoised voice recording; b) automatically generated transcripts without the denoising process; and c) manuscript transcripts. Denoising was implemented using noisereduce, a Python library to minimize background noise (CITATION). Automatic transcripts were produced using two toolkits: CMUSpinx pocketsphinx interpreter (PS) and Mozilla DeepSpeech (DS). The Small‐Bidirectional Encoder Representations from Transformers (BERT) model was implemented to extract high dimensional feature vectors to provide contextual representation of transcripts for classification analysis. The experiment was designed to compare the performance of transcript type (manual vs. automatic), recording quality (noise vs. denoise), as well as speech recognition methods (DS vs. PS). A neural network model was constructed to classify between those with AD and healthy controls. The classification analysis was conducted using 10‐fold cross validation. The classification was repeated 20 times and the averaged results were reported (see Table 1).ResultThe performance of each approach is shown as Table 1. Overall, the model that used the features extracted from DS transcripts with denoised speech recordings achieved the best performance which had 92.72% accuracy for AD diagnosis validation results. The comparison between raw and denoised speech recordings showed conflicting performances. Using features from PS transcripts with noisy speech recordings obtained better performance than the denoised one, which was different with features from DS transcripts.ConclusionThe analysis approach which combined transcription driven recognition and transfer learning indicated the potential of better performance in AD diagnosis. Further investigation is needed to understand the reason for its superior performance.

Reconstruction of a Fully Paralleled Auditory Spiking Neural Network and FPGA Implementation.

This paper presents a field-programmable gate array (FPGA) implementation of an auditory system, which is biologically inspired and has the advantages of robustness and anti-noise ability. We propose an FPGA implementation of an eleven-channel hierarchical spiking neuron network (SNN) model, which has a sparsely connected architecture with low power consumption. According to the mechanism of the auditory pathway in human brain, spiking trains generated by the cochlea are analyzed in the hierarchical SNN, and the specific word can be identified by a Bayesian classifier. Modified leaky integrate-and-fire (LIF) model is used to realize the hierarchical SNN, which achieves both high efficiency and low hardware consumption. The hierarchical SNN implemented on FPGA enables the auditory system to be operated at high speed and can be interfaced and applied with external machines and sensors. A set of speech from different speakers mixed with noise are used as input to test the performance our system, and the experimental results show that the system can classify words in a biologically plausible way with the presence of noise. The method of our system is flexible and the system can be modified into desirable scale. These confirm that the proposed biologically plausible auditory system provides a better method for on-chip speech recognition. Compare to the state-of-the-art, our auditory system achieves a higher speed with a maximum frequency of 65.03 MHz and a lower energy consumption of 276.83 μJ for a single operation. It can be applied in the field of brain-computer interface and intelligent robots.

Implementation of Embedded Unspecific Continuous English Speech Recognition Based on HMM

Background: To improve the usage of computers, a very important role is played by the Human-computer interaction through Natural Language Conversational Interfaces. Speech recognition technology allows the machine to understand human language. To achieve this function, a speech recognition algorithm is used. Methods: In order to realize the embedded speech recognition function based on HMM under the ARM platform, this paper, mainly based on the basic theoretical research of speech signals, establishes the HMM model, uses speech collection, recognition and other methods, simulates on MATLAB, and integrates the recognition system ported to ARM for debugging and running. Results: The research results show that the accuracy of HMM model experimental recognition can reach 98%, and the speed of speech recognition simulation on ARM is faster than speech recognition. The innovation of this paper is to make a comprehensive system introduction to embedded speech recognition from the perspective of embedded systems, and perform speech recognition simulation on MATLAB. Conclusion: The conclusion shows that the HMM-based embedded unspecific continuous English speech recognition system has high recognition accuracy and fast speed.

Application of Transfer Learning Algorithm and Real Time Speech Detection in Music Education Platform

Artificial intelligence (AI), particularly machine learning (ML) and neural networks (NN), has various applications and has sparked a lot of interest in the recent years due to its superior performance in a variety of tasks. Automatic speech recognition (ASR) is a technique that is becoming more important with the passage of time and is being used in our daily lives. Speech recognition is an important application of ML and NN, which is the auditory system of machines that realize the communication between humans and machines. In general, speech recognition methods are divided into three types, i.e., based on the channel model and speech knowledge method, template matching scheme, and the use of NN method. The main problem associated with the existing speech recognition methods is the low recognition accuracy and more computation time. In order to overcome the problem of low recognition accuracy of existing speech recognition techniques, a speech recognition technology based on the combination of deep convolution neural network (DCNN) algorithm and transfer learning techniques, i.e., VGG-16, is proposed in this study. Due to the limited application range of DCNN, when the input and output parameters are changed, it is necessary to reconstruct the model that leads to a long training time of the architecture. Therefore, the migration learning method is conducive to reducing the size of the dataset. Various experiments have been performed using different dataset constructs. The simulation results show that transfer learning is not only suitable for the comparison between the source dataset and the target dataset, but also suitable for two different datasets. The application of small datasets not only reduces the time and cost of dataset generation, but also reduces the training time and the requirement of computing power. From the experimental results, it is quite obvious that the proposed system performed better than the existing speech recognition methods, and its performance is superior in terms of recognition accuracy than the other approaches.

An Improved Speech Base Frequency Extraction Algorithm

Traditional fundamental frequency extraction algorithms mainly include cepstrum method, short-time autocorrelation method. Some end-point extraction methods can’t correctly select the part of effective value due to the particularity of dialect audio and the noise influence. And, some fundamental frequency extraction methods cannot get the data consistent with the trend of tuning values. Against the above, this paper proposes an improved double threshold method for the extraction of the basic frequency according to the characteristics of the vowels of the initials of the one-word tone, which extracts the continuous vowels of the one-word tone, and looks for the basic frequency extraction algorithm that meets the classification of the tones of one-word. Finally, the effective basic frequency data is obtained. This method is more effective than other methods in speech tone recognition. It is of great significance to the study of speech tone, especially the study of dialect tone.

Gsdnet: Gated Self-Supervised Denoising Speech Control Network

The cost of labeling data remains high, even with the effective implementation of deep neural networks in speech recognition. At the same time, noise still hampers the performance of speech-recognition methods. Thus, it is still challenging to make full use of data sets to enhance the robustness of recognition systems. In this letter, we construct GSDNet, a gated self-supervised denoising speech control network that consists of three parts (a denoising feature-extraction frontend, a speech recognition encoder, and a decoder based on gated convolutionary neural networks with self-supervised regression), to provide a low-cost method for training a robust speech recognition system, and we apply it to equipment-control tasks. Finally, the experimental results with the THCH30 and AISHELL data sets for equipment control show that the word error rate is less than 0.2 without a language model.

Psychosis speech recognition algorithm based on deep embedded sparse stacked autoencoder and manifold ensemble

Speech feature learning is the core and key of speech recognition method for mental illness. Deep feature learning can automatically extract speech features, but it is limited by the problem of small samples. Traditional feature extraction (original features) can avoid the impact of small samples, but it relies heavily on experience and is poorly adaptive. To solve this problem, this paper proposes a deep embedded hybrid feature sparse stack autoencoder manifold ensemble algorithm. Firstly, based on the prior knowledge, the psychotic speech features are extracted, and the original features are constructed. Secondly, the original features are embedded in the sparse stack autoencoder (deep network), and the output of the hidden layer is filtered to enhance the complementarity between the deep features and the original features. Third, the L1 regularization feature selection mechanism is designed to compress the dimensions of the mixed feature set composed of deep features and original features. Finally, a weighted local preserving projection algorithm and an ensemble learning mechanism are designed, and a manifold projection classifier ensemble model is constructed, which further improves the classification stability of feature fusion under small samples. In addition, this paper designs a medium-to-large-scale psychotic speech collection program for the first time, collects and constructs a large-scale Chinese psychotic speech database for the verification of psychotic speech recognition algorithms. The experimental results show that the main innovation of the algorithm is effective, and the classification accuracy is better than other representative algorithms, and the maximum improvement is 3.3%. In conclusion, this paper proposes a new method of psychotic speech recognition based on embedded mixed sparse stack autoencoder and manifold ensemble, which effectively improves the recognition rate of psychotic speech.

Research on Speech Recognition Method in Multi Layer Perceptual Network Environment

Aiming at the problem of low recognition rate in speech recognition methods, a speech recognition method in multi-layer perceptual network environment is proposed. In the multi-layer perceptual network environment, the speech signal is processed in the filter by using the transfer function of the filter. According to the framing process, the speech signal is windowed and framing processed to remove the silence segment of the speech signal. At the same time, the average energy of the speech signal is calculated and the zero crossing rate is calculated to extract the characteristics of the speech signal. By analyzing the principle of speech signal recognition, the process of speech recognition is designed, and the speech recognition in multi-layer perceptual network environment is realized. The experimental results show that the speech recognition method designed in this paper has good speech recognition performance

Arabic speech recognition using end‐to‐end deep learning

Arabic automatic speech recognition (ASR) methods with diacritics have the ability to be integrated with other systems better than Arabic ASR methods without diacritics. In this work, the application of state-of-the-art end-to-end deep learning approaches is investigated to build a robust diacritised Arabic ASR. These approaches are based on the Mel-Frequency Cepstral Coefficients and the log Mel-Scale Filter Bank energies as acoustic features. To the best of our knowledge, end-to-end deep learning approach has not been used in the task of diacritised Arabic automatic speech recognition. To fill this gap, this work presents a new CTC-based ASR, CNN-LSTM, and an attention-based end-to-end approach for improving diacritisedArabic ASR. In addition, a word-based language model is employed to achieve better results. The end-to-end approaches applied in this work are based on state-of-the-art frameworks, namely ESPnet and Espresso. Training and testing of these frameworks are performed based on the Standard Arabic Single Speaker Corpus (SASSC), which contains 7 h of modern standard Arabic speech. Experimental results show that the CNN-LSTM with an attention framework outperforms conventional ASR and the Joint CTC-attention ASR framework in the task of Arabic speech recognition. The CNN-LSTM with an attention framework could achieve a word error rate better than conventional ASR and the Joint CTC-attention ASR by 5.24% and 2.62%, respectively.

Environmental effects on reliability and accuracy of MFCC based voice recognition for industrial human-robot-interaction

Collaborative robots are becoming increasingly important for advanced manufacturing processes. The purpose of this paper is to determine the capability of a novel Human-Robot-interface to be used for machine hole drilling. Using a developed voice activation system, environmental factors on speech recognition accuracy are considered. The research investigates the accuracy of a Mel Frequency Cepstral Coefficients-based feature extraction algorithm which uses Dynamic Time Warping to compare an utterance to a limited, user-dependent dictionary. The developed Speech Recognition method allows for Human-Robot-Interaction using a novel integration method between the voice recognition and robot. The system can be utilised in many manufacturing environments where robot motions can be coupled to voice inputs rather than using time consuming physical interfaces. However, there are limitations to uptake in industries where the volume of background machine noise is high.

Voiceless Speech Recognition System

Speech recognition is a very important part of the human-machine interaction system. The commonly used methods for speech recognition are invariantly based on the analysis of speech signals. These signals might not be always viable especially in noisy environments or in environments where discretion is of utmost importance. The “Thought Predictor system” allows the user to converse silently with a computing device. It thus enables the user to communicate with devices and AI assistants in a voiceless and discrete manner. The user’s intention to speak and internal speech is characterized by neuromuscular signals by the speech articulating muscles that are captured by this system to reconstruct this speech.