An Efficient Speech Recognition System

In this paper, based on the study of two speech recognition algorithms, two designs of speech recognition system are given to realize this isolated speech recognition mobile robot control system based on ARM9 processor. The speech recognition process includes pretreatment of speech signal, characteristic extrication, pattern matching and post-processing. Mel-Frequency cepstrum coefficients (MFCC) and linear prediction cepstrum coefficients (LPCC) are the two most common parameters. Through analysis and comparison the parameters, MFCC shows more noise immunity than LPCC, so MFCC is selected as the characteristic parameters. Both dynamic time warping (DTW) and hidden markov model (HMM) are commonly used algorithm. For the different characteristics of DTW and HMM recognition algorithm, two different programs were designed for mobile robot control system. The effect and speed of the two speech recognition system were analyzed and compared.

As one of biometric identification technologies, speaker recognition shows better application prospects in many fields. At present, the implementation of speaker recognition algorithm on the hardware is mostly based on System on a Programmable Chip(SOPC) platform of Field Programmable Gate Array(FPGA) with Nios II Intellectual Property(IP) core. And the algorithm can be selected and optimized effectively on SOPC. However the high-speed and parallel operation of FPGA can't be fully utilized. By researching and analyzing the speaker recognition algorithm, a FPGA-based speaker recognition method is presented in this paper. This method includes the Voice Active Detection (VAD), Mel Frequency Cepstrum Coefficient (MFCC) extraction and Vector Quantization (VQ) recognition algorithm. By using the technology such as Ping-pang operation, Pipeline and module reuse, it makes full use of FPGA's speed. After testing, this method can effectively meet the requirement of real-time data processing.

An audio-visual speech recognition system (AVSR) integrates audio and visual information to perform speech recognition task. The AVSR has various applications in practice especially in natural language processing systems such as speech-to-text conversion, automatic translation and sentiment analysis. Decades ago, researchers tend to use Hidden Markov Model (HMM) to construct speech recognition system due to its good achievements in success recognition rate. However, HMM's training dataset is enormous in order to have sufficient linguistic coverage. Besides, its recognition rate under noisy environments is not satisfying. To overcome this deficiency, a Recurrent Neural Network (RNN) based AVSR is proposed. The proposed AVSR model consists of three components: 1) audio features extraction mechanism, 2) visual features extraction mechanism and 3) audio and visual features integration mechanism. The features integration mechanism combines the output features from both audio and visual extraction mechanisms to generate final classification results. In this research, the audio features mechanism is modelled by Mel-frequency Cepstrum Coefficient (MFCC) and further processed by RNN system, whereas the visual features mechanism is modelled by Haar-Cascade Detection with OpenCV and again, it is further processed by RNN system. Then, both of these extracted features were integrated by multimodal RNN-based features-integration mechanism. The performance in terms of the speech recognition rate and the robustness of the proposed AVSR system were evaluated using speech under clean environment and Signal Noise Ratio (SNR) levels ranging from -20 dB to 20 dB with 5 dB interval. On average, final speech recognition rate is 89% across different levels of SNR.

In recent years, crime prevention systems have been developed to detect various hazardous situations. In general, the systems utilize the image information recorded by a camera to monitor the situations. It is however difficult to detect them in the blind area. To address the problem, it is required to utilize not only image information but also acoustic information occurred in such situations. Our previous study showed that two acoustic features including rahmonic and mel-frequency cepstrum coefficients (MFCCs) are effective for detecting the shouted speech. Rahmonic shows a subharmonic of fundamental frequency in the cepstrum domain, and MFCCs represent coefficients that collectively make up mel-frequency cepstrum. In this method, a shouted speech model is constructed from these features by using a gaussian mixture model (GMM). However, the previous method with GMM has difficulty in representing temporal changes of the speech features. In this study, we further expand the previous method using hidden Markov model (HMM) which has state transition to represent the temporal changes. Through objective experiments, the proposed method using HMM could achieve higher detection performance of the shouted speech than the conventional method using GMM.

ABSTRACT: Speech is one of the most natural and efficient forms of communication among humans, and it is becoming increasingly integral to human-computer interaction in modern applications such as virtual assistants, smart devices, and automated customer service systems. Among the various technologies that leverage speech, speaker recognition stands out as a biometric method that identifies or verifies individuals based on the unique characteristics of their voice. This approach has gained significant traction due to its non-intrusive nature, ease of integration, and applicability across various domains such as security, forensics, and personalised user experiences. In this paper, we introduce a novel speaker recognition system specifically designed to identify speakers based on utterances in the Marathi language, a linguistically rich and widely spoken regional language in India. The system utilises Mel-Frequency Cepstral Coefficients (MFCCS) to extract distinguishing vocal features that closely mimic the human auditory perception. MFCCS are particularly effective in capturing speech’s phonetic and acoustic properties, making them a preferred choice in speech and speaker recognition tasks. Vector quantisation (VQ) is applied to optimise the feature set and reduce computational complexity. VQ compresses the high-dimensional MFCC feature vectors into representative clusters without significantly compromising accuracy, thereby enhancing system efficiency. These features are then processed using hidden Markov Models (HMMS), which are adept at modelling temporal sequences and dynamic variations in speech. HMMS offer a statistical framework that effectively captures the sequential nature of speech patterns, leading to more reliable speaker modelling and recognition.

Digit speech recognition is important in many applications such as automatic data entry, PIN entry, voice dialing telephone, automated banking system, etc. This paper presents speaker independent speech recognition system for Malayalam digits. The system employs Mel frequency cepstrum coefficient (MFCC) as feature for signal processing and Hidden Markov model (HMM) for recognition. The system is trained with 21 male and female voices in the age group of 20 to 40 years and there was 98.5% word recognition accuracy (94.8% sentence recognition accuracy) on a test set of continuous digit recognition task.

Automatic Speech Recognition has been an active topic of research for the past four decades. The main objective of the automatic speech recognition task is to convert a speech segment into an interpretable text message without the need of human intervention. Many different algorithms and schemes based on different mathematical paradigms have been proposed in an attempt to improve recognition rates. Cepstral coefficients play an important part in speech theory and in automatic speech recognition in particular due to their ability to compactly represent relevant information that is contained in a short time sample of a continuous speech signal. The goal of this paper is to discuss comparison of speech parameterization methods: Mel-Frequency Cepstrum Coefficients (MFCC) and improved Mel-Frequency Cepstrum Coefficients (MFCC) using RASTA filters. Thus, in this study, we try to improve the MFCC algorithms to achieve much accuracy reducing the error rates in Automatic Speech Recognition. First, we remove signal correlation through normalization, then we use RASTA filter to filtering the cepstral coefficients. Finally, we reduce dimension of the cepstral coefficients by the variances of cepstral coefficients in different dimension and obtain our features. By using various classifiers, we try to simulate the speech feature extraction at much optimal and least error rate providing robust method for Automatic Speech Recognition (ASRs).KeywordsAutomatic Speech RecognitionMel frequency Cepstrum Coefficients (MFCC’s)ERB Gammatone FilteringHidden Markov Model

This paper presents a study on the effectiveness of mel-frequency cepstrum coefficients (MFCCs) and some of their statistical distribution properties (skewness, kurtosis, standard deviation) as the features for text-dependent speaker identification. Multi-layer neural network with backpropagation learning algorithm is used here as the classification tool. The MFCCs representing the speaker characteristics of a speech segment are computed by nonlinear filterbank analysis and discrete cosine transform. The speaker identification efficiency and the convergence speed of the neural network are investigated for different combinations of the proposed features. The result shows that the first MFCC degrades the identification competence and statistical distribution parameters enhance the training speed of the neural network.

A model-based framework of classification error rate estimation is proposed for speech and speaker recognition. It aims at predicting the run-time performance of a hidden Markov model (HMM) based recognition system for a given task vocabulary and grammar without the need of running recognition experiments using a separate set of testing samples. This is highly desirable both in theory and in practice. However, the error rate expression in HMM-based speech recognition systems has no closed form solution due to the complexity of the multi-class comparison process and the need for dynamic time warping to handle various speech patterns. To alleviate the difficulty, we propose a one-dimensional model-based misclassification measure to evaluate the distance between a particular model of interest and a combination of many of its competing models. The error rate for a class characterized by the HMM is then the value of a smoothed zero-one error function given the misclassification measure. The overall error rate of the task vocabulary could then be computed as a function of all the available class error rates. The key here is to evaluate the misclassification measure in terms of the parameters of environmental-matched models without running recognition experiments, where the models are adapted by very limited data that could be just the testing utterance itself. In this paper, we show how the misclassification measure could be approximated by first computing the distance between two mixture Gaussian densities, then between two HMMs with mixture Gaussian state observation densities and finally between two sequences of HMMs. The misclassification measure is then converted into classification error rate. When comparing the error rate obtained in running actual experiments and that of the new framework, the proposed algorithm accurately estimates the classification error rate for many types of speech and speaker recognition problems. Based on the same framework, it is also demonstrated that the error rate of a recognition system in a noisy environment could also be predicted.

The model-based speech enhancement method usually models types of noises in a prior and selects one noise model in the enhancement phase. In this paper, we study the modeling and selection (i.e. noise identification) algorithm for the model-base speech enhancement. For the noise signal features, Mel Frequency Cepstrum Coefficient (MFCC), Liner Prediction Coefficient (LPC), Linear Spectral Frequency (LSF) and Linear Predictive Cepstrum Coefficients (LPCC) are utilized. For the noise model, codebook, Gaussian Mixture Model (GMM) and Hidden Markov Model (HMM) are adopted. We couple these features and models and provide a comparative performance analysis of them. In the experiment, we use NOIZEUS database for noise modeling and testing. The experiment results show that MFCC and GMM is an excellent couple for noise identification, and LFS can take the place of LPC in the codebook-based speech enhancement.

Computer Assisted Pronunciation Training (CAPT) systems aim to provide immediate, individualized feedback to the user on the overall quality of the pronunciation made. In such systems, one must be able to extract features from a waveform and represent words in the vocabulary. This paper presents the performance of Hidden Markov Model (HMM), Support-Vector Machine (SVM) and Multilayer Perceptron (MLP) as automatic speech recognizers for the English digits spoken by Filipino speakers. Speech waveforms are translated into a set of feature vectors using Mel Frequency Cepstrum Coefficients (MFCC). The training set consists of speech samples recorded by native Filipinos who speak English. The HMM-trained model produced a recognition rate of 95.79% compared to 86.33% and 91.66% recognition rates of SVM and MLP, respectively <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> .

&lt;p&gt;In Ethiopia, the largest ethnic and linguistic groups are the Oromos, Amharas and Tigrayans. This paper presents the performance analysis of text-independent speaker identification system for the Amharic language in noisy environments. VQ (Vector Quantization), GMM (Gaussian Mixture Models), BPNN (Back propagation neural network), MFCC (Mel-frequency cepstrum coefficients), GFCC (Gammatone Frequency Cepstral Coefficients), and a hybrid approach had been use as techniques for identifying speakers of Amharic language in noisy environments. For the identification process, speech signals are collected from different speakers including both sexes; for our data set, a total of 90 speakers’ speech samples were collected, and each speech have 10 seconds duration from each individual. From these speakers, 59.2%, 70.9% and 84.7% accuracy are achieved when VQ, GMM and BPNN are used on the combined feature vector of MFCC and GFCC. &lt;/p&gt;

This study analyzes the effect of stress in human and automatic stressed speech processing tasks for speech collected from non-professional speakers. The database of 33 keywords is collected under five stress conditions, namely, neutral, angry, happy, sad and Lombard from fifteen speakers. The first study is to understand the ability to identify stress by human and automatic speech processing. The average performance of human stress classification is 59.44%. The average performance of automatic stress classifier using Vector Quantization (VQ) and Hidden Markov model (HMM) is 54.65% and 56.02%, respectively. The second study has been done to understand the effect of stress in human and automatic speech recognition. The average performance of human stressed speech recognition is 99.60%. The automatic stressed speech recognition performance using VQ and HMM is 82.42% and 76.79%, respectively. Even in the non-professional speakers scenario, human performance is better than automatic processing. Also, automatic processing seem to show considerable degradation in performance that warrant development of new methods to handle stress information.

Our goal is to realize a humanoid robot that has the capabilities of recognizing simultaneous speech. A humanoid robot under real-world environments usually hears a mixture of sounds, and thus three capabilities are essential for robot audition; sound source localization, separation, and recognition of separated sounds. In particular, an interface between sound source separation and speech recognition is important. In this paper, we designed an interface between sound source separation and speech recogniton by applying Missing Feature Theory (MFT) . In this method, spectral sub-bands distorted by sound source separation are detected from input speech as missing features. The detected missing features are masked on recognition not to affect the system badly. Therefore, this method is more flexible when noises change dynamically and drastically. It is the most important issue how distorted spectral sub-bands are detected. To solve the issue, we used speech feature apropriate for MFT-based ASR, and developed automatic missing feature mask generation. As a speech feature, we used a Mel-Scale Log Spectral (MSLS) feature instead of Mel-Frequency Cepstrum Coefficient (MFCC) which is commonly used for ASR. We presented a method of generating missing feature mask automatically by using information from sound source separation. To evaluate our method, we implemented it in a humanoid robotSIG2, and performed the experiments on recognition of three simultaneous isolated words. As a result, our method outperformed conventional ASR with MSLS feature.

Voice activation and speaker recognition are needed in many applications today. Speaker recognition is the process of automatically recognizing who speaks based on the voice signal. The introduction of these speakers is generally required on systems that use security and privacy. One example of this paper application is for activation and security in controlling humanoid robots. Voice recording process using Kinect 2.0. The first step in the speech recognition process is feature extraction. In this paper use Mel Frequency Cepstrum Coefficient (MFCC) on characteristic extraction process and Dynamic Time Warping (DTW) used as feature matching technique. The test was performed by 5 different speakers, with 2 types of words (“aktifkan” means activate and “hello slim”), and test with different recording distance (0.5m, 2m, 4m). Robot activation using two different types of words has an average accuracy of 91.5%. At the next difficulty level for testing the recording distance accuracy decreased from 97.5% to 85% to 65%.