• 전자공학회논문지B
• /
• 제32B권5호
• /
• pp.756-764
• /
• 1995

In this paper, we proposed MLP(MultiLayer Perceptron) based HMM's(Hidden Markov Models) with bounded state duration for isolated word recognition. The minimum and maximum state duration for each state of a HMM are estimated during the training phase and used as parameters of constraining state transition in a recognition phase. The procedure for estimating these parameters and the recognition algorithm using the proposed HMM's are also described. Speaker independent isolated word recognition experiments using a vocabulary of 10 city names and 11 digits indicate that recognition rate can be improved by adjusting the minimum state durations.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1998년도 추계종합학술대회 논문집
• /
• pp.1255-1258
• /
• 1998

Hidden Markov Model (HMM) is the most widely used method in speech recognition. In general, HMM parameters are trained to have maximum likelihood (ML) for training data. This method doesn't take account of discrimination to other words. To complement this problem, this paper proposes a word verification method by re-recognition of the recognized word and its similar word using the discriminative function between two words. The similar word is selected by calculating the probability of other words to each HMM. The recognizer haveing discrimination to each word is realized using the weighting to each state and the weighting is calculated by genetic algorithm.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2009년도 정보 및 제어 심포지움 논문집
• /
• pp.37-39
• /
• 2009

In the research of speech recognition, locating the beginning and end of a speech utterance in a background of noise is of great importance. Since the background noise presenting to record will introduce disturbance while we just want to get the stationary parameters to represent the corresponding speech section, in particular, a major source of error in automatic recognition system of isolated words is the inaccurate detection of beginning and ending boundaries of test and reference templates, thus we must find potent method to remove the unnecessary regions of a speech signal. The conventional methods for speech endpoint detection are based on two simple time-domain measurements - short-time energy, and short-time zero-crossing rate, which couldn't guarantee the precise results if in the low signal-to-noise ratio environments. This paper proposes a novel approach that finds the Lyapunov exponent of time-domain waveform. This proposed method has no use for obtaining the frequency-domain parameters for endpoint detection process, e.g. Mel-Scale Features, which have been introduced in other paper. Comparing with the conventional methods based on short-time energy and short-time zero-crossing rate, the novel approach based on time-domain Lyapunov Exponents(LEs) is low complexity and suitable for Digital Isolated Word Recognition System.

• 전자공학회논문지B
• /
• 제32B권11호
• /
• pp.1527-1535
• /
• 1995

This paper introduces an isolated word recognition system realized on a neurocomputer called E-MIND II, which is a 2-D torus wavefront array processor consisting of 256 DNP IIs. The DNP II is an all digital VLSI unit processor for the EMIND II featuring the emulation capability of more than thousands of neurons, the 40 MHz clock speed, and the on-chip learning. Built by these PEs in 2-D toroidal mesh architecture, the E- MIND II can be accelerated over 2 Gcps computation speed. In this light, the advantages of the E-MIND II in its capability of computing speed, scalability, computer interface, and learning are especially suitable for real time application such as speech recognition. We show how to map a TDNN structure on this array and how to code the learning and recognition algorithms for a user independent isolated word recognition. Through hardware simulation, we show that recognition rate of this system is about 97% for 30 command words for a robot control.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1988년도 전기.전자공학 학술대회 논문집
• /
• pp.10-13
• /
• 1988

An isolated-word recognition method using adaptively partitioned multisection codebooks is proposed. Each training utterance was divided into several sections according to its pattern extracted by labeling technique. For each pattern, reference codebooks were generated by clustering the training vectors of the same section. In recognition procedure, input speech was divided into the sections by the same method used in codebook generation procedure, and recognized to the reference word whose codebook represented the smallest average distortion. The proposed method was tested for 100 Korean words and attained recognition rate about 96 percent.

• 한국음향학회지
• /
• 제16권1호
• /
• pp.96-99
• /
• 1997

본 논문에서는 고립단어 인식을 위한 빠른 전처리기를 소개한다. 제안하는 전처리기는 적은 계산량으로 후보 단어를 추출한다. 본 전처리기에서는 계산량을 줄이기 위해서 벡터 양자화 대신에 특징 정렬 알고리즘을 사용하였다. 이 전처리기의 유효성을 보이기 위해서 준연속 은닉 마코프 모델을 기반으로 한 음성 인식기와 벡터 양자화를 기반으로 한 전처리기에 대해서 화자독립 고립단어 인식에 대한 성능을 비교했다. 실험에 사용한 음성 데이터는 남성 호자 40명이 발성한 244 단어이며, 40명의 화자 중에서 20명은 전처리기의 훈련용으로 사용했으며 나머지 20명은 평가용으로 사용하였다. 실험의 결과, 음성 데이터에 대해서 90%의 감축을 조건에서 제안한 전처리기는 99.9%의 정확성을 보였다.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1994년도 FIFTH WESTERN PACIFIC REGIONAL ACOUSTICS CONFERENCE SEOUL KOREA
• /
• pp.1039-1045
• /
• 1994

The speech recognition systems using VQ have usually the problem decreasing recognition rate, MSVQ assigning the dissimilar vectors to a segment. In this paper, applying One-stage DMS/DP algorithm to the recognition experiments, we can solve these problems to what degree. Recognition experiment is peformed for Korean DDD area names with DMS model of 20 sections and word unit template. We carried out the experiment in speaker dependent and speaker independent, and get a recognition rates of 97.7% and 81.7% respectively.

• 대한전자공학회논문지
• /
• 제25권4호
• /
• pp.356-365
• /
• 1988

In this paper, we describe an isolated words recognition system for Korean isolated digits based on a voiced -unvoiced decision algorithm and a frequency domain analysis. The algorithm first performs a voiced-unvoiced decision procedure for the begtinning part of each uttered work using the normalized log energy and zero crossing rate as decision parameters. Based on this decision,. each word is assigned to one of two classes. In order to identify the uttered word within each class, a dynamic time warping algorithm is applied using formant frequencies as the basis for the distance measure. We exploit a pole-zero analysis to measure formant frequencies in each frame. We have observed that pole-zero analysis can provide more accurate estimation of formant frequencies than analysis based on poles only. Experimental recognition rates of 97.3% illustrating the performance of the recognition system was achieved.

• 한국음향학회지
• /
• 제16권6호
• /
• pp.106-109
• /
• 1997

고립단어 음성인식 시스템이 실용적이 되려면 인식 대상 이외의 단어를 거절할 수 있는 기능이 요구된다. 본 논문에서는 집단화된 음소 모델과 likelihood ratio에 의한 후처리 방법을 사용하여 거절기능을 구현하는 방법을 제안하였다. 기본적인 음성인식 시스템은 단어 단위 연속 HMM을 사용하였고, 6개의 집단화된 음소 모델들은 음성학적으로 균형잡힌 음성 데이터베이스를 이용하여 훈련된 45개의 문맥독립 음소 모델들로부터 통계적 방법에 의하여 생성되었다. 22개의 부서 명칭을 대상으로 한 화자독립 고립단어 인식시스템에서 거절성능을 시험하여 본 결과, 가장 높은 확률값과 두 번째 높은 확률값을 가지는 후보단어들 간의 차이값에 의하여 거절기능을 수행하는 기존의 후처리 방법보다 성능이 향상됨을 알 수 있었다. 또한 이 집단화된 음소모델은 인식 대상 어휘가 다른 고립단어 인식 시스템에도 재훈련 없이 그대로 사용될 수 있다.

• 한국정보처리학회논문지
• /
• 제7권6호
• /
• pp.1964-1972
• /
• 2000

For the Korean language, on acoustics, it creates a different form of phonetic value not a phoneme by its own peculiar property. Therefore, the construction of extended recognition system for understanding Korean language should be created with a study of the Korean rule-based system, before it can be used as post-processing of the Korean recognition system. In this paper, text-based Korean rule-based system featuring Korean peculiar vocal sound changing rule is constructed. and based on the text-based phonetic value result of the system constructed, a preliminary phonetic value segmentation border points with non-uniform blocks are extracted in Korean isolated word speech. Through the way of merge and recognition of the non-uniform blocks between the extracted border points, recognition possibility of Korean voice as the form of the phonetic vale has been investigated.