• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.312-318
• /
• 2022

This paper proposes a model which classifies the type of guns and information about sound source location using deep neural network. The proposed classification model is composed of convolutional neural networks (CNN) and long short-term memory (LSTM). For training and test the model, we use the Gunshot Audio Forensic Dataset generated by the project supported by the National Institute of Justice (NIJ). The acoustic signals are transformed to Mel-Spectrogram and they are provided as learning and test data for the proposed model. The model is compared with the control model consisting of convolutional neural networks only. The proposed model shows high accuracy more than 90 %.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.1
• /
• pp.373-377
• /
• 2017

This paper presents a voice-activity detection (VAD) method for sound sequences with various SNRs. For real-time VAD applications, it is inadequate to employ a post-processing for the removal of burst clippings from the VAD output decision. To tackle this problem, building on the bi-level hidden Markov model, for which a state layer is inserted into a typical hidden Markov model (HMM), we formulated a robust method for VAD not requiring any additional post-processing. In the method, a forward-inference-ratio test was devised to detect the speech endpoints and Mel-frequency cepstral coefficients (MFCC) were used as the features. Our experiment results show that, regarding different SNRs, the performance of the proposed approach is more outstanding than those of the conventional methods.

• Journal of Intelligence and Information Systems
• /
• v.29 no.3
• /
• pp.57-78
• /
• 2023

COVID-19, which started in Wuhan, China in November 2019, spread beyond China in 2020 and spread worldwide in March 2020. It is important to prevent a highly contagious virus like COVID-19 in advance and to actively treat it when confirmed, but it is more important to identify the confirmed fact quickly and prevent its spread since it is a virus that spreads quickly. However, PCR test to check for infection is costly and time consuming, and self-kit test is also easy to access, but the cost of the kit is not easy to receive every time. Therefore, if it is possible to determine whether or not a person is positive for COVID-19 based on the sound of a cough so that anyone can use it easily, anyone can easily check whether or not they are confirmed at anytime, anywhere, and it can have great economic advantages. In this study, an experiment was conducted on a method to identify whether or not COVID-19 was confirmed based on a cough sound. Cough sound features were extracted through MFCC, Mel-Spectrogram, and spectral contrast. For the quality of cough sound, noisy data was deleted through SNR, and only the cough sound was extracted from the voice file through chunk. Since the objective is COVID-19 positive and negative classification, learning was performed through XGBoost, LightGBM, and FCNN algorithms, which are often used for classification, and the results were compared. Additionally, we conducted a comparative experiment on the performance of the model using multidimensional vectors obtained by converting cough sounds into both images and vectors. The experimental results showed that the LightGBM model utilizing features obtained by converting basic information about health status and cough sounds into multidimensional vectors through MFCC, Mel-Spectogram, Spectral contrast, and Spectrogram achieved the highest accuracy of 0.74.

• IEIE Transactions on Smart Processing and Computing
• /
• v.3 no.5
• /
• pp.259-266
• /
• 2014

In this paper, we propose a novel technique for noise robust automatic speech recognition (ASR). The development of ASR techniques has made it possible to recognize isolated words with a near perfect word recognition rate. However, in a highly noisy environment, a distinct mismatch between the trained speech and the test data results in a significantly degraded word recognition rate (WRA). Unlike conventional ASR systems employing Mel-frequency cepstral coefficients (MFCCs) and a hidden Markov model (HMM), this study employ histogram of oriented gradient (HOG) features and a Support Vector Machine (SVM) to ASR tasks to overcome this problem. Our proposed ASR system is less vulnerable to external interference noise, and achieves a higher WRA compared to a conventional ASR system equipped with MFCCs and an HMM. The performance of our proposed ASR system was evaluated using a phonetically balanced word (PBW) set mixed with artificially added noise.

• MALSORI
• /
• no.53
• /
• pp.129-141
• /
• 2005

The goal of this research is to improve the quality of reconstructed speech in the Distributed Speech Recognition (DSR) system. For the extended DSR, we estimate the variable Maximum Voiced Frequency (MVF) from Mel-Frequency Cepstral Coefficient (MFCC) based on Gaussian Mixture Model (GMM), to implement realistic harmonic plus noise model for the excitation signal. For the standard DSR, we also make the voiced/unvoiced decision from MFCC based on GMM because the pitch information is not available in that case. The perceptual test reveals that speech reconstructed by the proposed method is preferred to the one by the conventional methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.8
• /
• pp.3257-3269
• /
• 2020

Articulation disorders are characterized by an inability to achieve clear pronunciation due to misuse of the articulators. In this paper, a method of detecting such disorders by comparing to the standard pronunciations is proposed. This method defines the standard pronunciations from the speeches of normal children by clustering them with three features which are the Linear Predictive Cepstral Coefficient (LPCC), the Mel-Frequency Cepstral Coefficient (MFCC), and the Relative Spectral Analysis Perceptual Linear Prediction (RASTA-PLP). By calculating the distance between the centroid of the standard pronunciation and the inputted pronunciation, disordered speech whose features locates outside the cluster is detected. 89 children (58 of normal children and 31 of children with disorders) were recruited. 35 U-TAP test words were selected and each word's standard pronunciation is made from normal children and compared to each pronunciation of children with disorders. In the experiments, the pronunciations with disorders were successfully distinguished from the standard pronunciations.

• Phonetics and Speech Sciences
• /
• v.2 no.1
• /
• pp.87-95
• /
• 2010

In this paper, we address issues associated with mixed-lingual text-to-speech based on context-dependent HMMs, where there are multiple sets of HMMs corresponding to each individual language. In particular, we propose smoothing techniques of synthesis parameters at the boundaries between different languages to obtain more natural quality of speech. In other words, mel-frequency cepstral coefficients (MFCCs) at the language boundaries are smoothed by applying several linear and nonlinear approximation techniques. It is shown from an informal listening test that synthesized speech smoothed by a modified version of linear least square approximation (MLLSA) and a quadratic interpolation (QI) method is preferred than that without using any smoothing technique.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06d
• /
• pp.60-63
• /
• 1998

본 논문에서는 사용자가 별도의 훈련과정 없이 인식대상 어휘를 추가 및 변경이 가능한 가변어휘 인식시스템에 관하여 기술한다. 가변어휘 음성인식에서는 미리 구성된 음소모델을 토대로 인식대상 어휘가 결정되명 발음사전에 의거하여 이들 어휘에 해당하는 음소모델을 연결함으로써 단어모델을 만든다. 사용된 음소모델은 현재 음소의 앞뒤의 음소 context를 고려한 문맥종속형(Context-Dependent)음소모델인 triphone을 사용하였고, 연속확률분포를 가지는 Hidden Markov Model(HMM)기반의 고립단어인식 시스템을 구현하였다. 비교를 위해 문맥 독립형 음소모델인 monophone으로 인식실험을 병행하였다. 개발된 시스템은 음성특징벡터로 MFCC(Mel Frequency Cepstrum Coefficient)를 사용하였으며, test 환경에서 나타나지 않은 unseen triphone 문제를 해결하기 위하여 state-tying 방법중 음성학적 지식에 기반을 둔 tree-based clustering 기법을 도입하였다. 음소모델 훈련에는 ETRI에서 구축한 POW (Phonetically Optimized Words) 음성 데이터베이스(DB)[1]를 사용하였고, 어휘독립인식실험에는 POW DB와 관련없는 22개의 부서명을 50명이 발음한 총 1.100개의 고립단어 부서 DB[2]를 사용하였다. 인식실험결과 문맥독립형 음소모델이 88.6%를 보인데 비해 문맥종속형 음소모델은 96.2%의 더 나은 성능을 보였다.

• Speech Sciences
• /
• v.13 no.3
• /
• pp.7-18
• /
• 2006

Diagnosis of pathological voice is one of the important issues in biomedical applications of speech technology. This study focuses on the discrimination of voice disorder using HMM (Hidden Markov Model) for automatic detection between normal voice and vocal fold disorder voice. This is a non-intrusive, non-expensive and fully automated method using only a speech sample of the subject. Speech data from normal people and patients were collected. Mel-frequency filter cepstral coefficients (MFCCs) were modeled by HMM classifier. Different states (3 states, 5 states and 7 states), 3 mixtures and left to right HMMs were formed. This method gives an accuracy of 93.8% for train data and 91.7% for test data in the discrimination of normal and vocal fold disorder voice for sustained /a/.

• Proceedings of the KAIS Fall Conference
• /
• 2010.05a
• /
• pp.257-260
• /
• 2010

In this paper, we propose the computation reduction methods of LSP(Line spectrum pairs) transformation that is mainly used in CELP vocoders. In order to decrease the computational time in real root method the characteristic of four proposed algorithms is as the following. First, scheme to reduce the LSP transformation time uses mel scale. Developed the second scheme is the control of searching order by the distribution characteristic of LSP parameters. Third, scheme to reduce the LSP transformation time uses voice characteristics. Developed the fourth scheme is the control of searching interval and order by the distribution characteristic of LSP parameters. As a result of searching time, computational amount, transformed LSP parameters, SNR, MOS test, waveform of synthesized speech, spectrogram analysis, searching time is reduced about 37.5%, 46.21%, 46.3%, 51.29% in average, computational amount is reduced about 44.76%, 49.44%, 47.03%, 57.40%. But the transformed LSP parameters of the proposed methods were the same as those of real root method.