• Proceedings of the KSPS conference
• /
• 2002.11a
• /
• pp.179-182
• /
• 2002

The development of noise robust speech processing systems is becoming increasingly important as speech technology is currently widely applied in real world applications. Recently, to resolve such a noise problem, adaptive noise canceller(ANC) is frequently used, which is based upon adaptive filters. The adaptive recursive filters perform better than adaptive non-recursive filters due to the added poles, but the stability may be severely threatened. But these problems of adaptive recursive filters was solved by ACHARF algorithm. This paper presents a method which combines speaker verification system with ANC(Adaptive Noise Canceller) using the ACHARF algorithm. In the front-end stage, ANC is adopted to suppress the additive noise imposed on the speech signal. The results show that the performance of speaker verification system becomes better than before.

• MALSORI
• /
• no.65
• /
• pp.125-141
• /
• 2008

In this work, we propose an automatic word spacing system for Korean using conditional random fields (CRFs) with Korean features. We map a word spacing problem into a classification problem in our work. We build a basic system which uses CRFs and Eumjeol bigram. After then, we analyze the result of inner-test. We extend a basic system added by some Korean features which are Josa, Eomi and two head Eumjeols of word extracting from lexicon. From the results of experiment, we can see that the proposed method is better than previous methods. Additionally the proposed method will be able to use mobile and speech applications because of very small size of model.

• MALSORI
• /
• no.57
• /
• pp.165-174
• /
• 2006

In this paper we use a training algorithm, MCE (Minimum Classification Error), to improve the performance of a text-independent speaker identification system. The MCE training scheme takes account of possible competing speaker hypotheses and tries to reduce the probability of incorrect hypotheses. Experiments performed on a small set speaker identification task show that the discriminant training method using MCE can reduce identification errors by up to 54% over a baseline system trained using Bayesian adaptation to derive GMM (Gaussian Mixture Models) speaker models from a UBM (Universal Background Model).

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.287-289
• /
• 2005

This paper presents a hybrid neural network system which used a Self-Organizing Map and Multilayer Perceptron for the problem of Malay syllables speech recognition. The novel idea in this system is the usage of a two-dimension Self-organizing feature map as a sequential mapping function which transform the phonetic similarities or acoustic vector sequences of the speech frame into trajectories in a square matrix where elements take on binary values. This property simplifies the classification task. An MLP is then used to classify the trajectories that each syllable in the vocabulary corresponds to. The system performance was evaluated for recognition of 15 Malay common syllables. The overall performance of the recognizer showed to be 91.8%.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.1
• /
• pp.171-179
• /
• 1995

In many continuous speech recognition systems, phoneme is used as a basic recognition unit However, the coarticulation generated among neighboring phonemes makes difficult to recognize phonemes consistently. This paper proposes allophone as an alternative recognition unit. We have classified each phoneme into three different allophone groups by the location of phoneme within a syllable. For a recognition algorithm, time-delay neural network(TDNN) has been designed. To recognize all Korean allophones, TDNNs are constructed in modular fashion according to acoustic-phonetic features (e.g. voiced/unvoiced, the location of phoneme within a word). Each TDNN is trained independently, and then they are integrated hierarchically into a whole speech recognition system. In this study, we have experimented Korean plosives with phoneme-based recognition system and allophone-based recognition system. Experimental results show that allophone-based recognition is much less affected by the coarticulation.

• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.41-44
• /
• 2005

This study is about Chinese tone evaluation system for Korean learners using speech technology, Chinese prounciaion system consists of initials, finals and tones. Initials/finals are in segmental level and tones are in suprasegmental level. So different method could be used assessing Korean users' Chinese. Differ from segmental level recognition method, we chose pattern matching method in evaluating Chinese tones. Firstly we defined speakers' own speech range and produced standard tonal pattern according to speakers' own range. And then we compared input patterns of users with referring patterns.

• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.45-48
• /
• 2005

This study is about constructing L2 pronunciation correction system for L1 speakers using speech technology. Chinese pronunciation system consists of initials, finals and tones. Initials/finals are in segmental level and tones are in suprasegmental level. So different method could be used assessing Korean users' Chinese. The recognition rate of initials is 81.9% and that of finals is 68.7% in the standard acoustic model. Differ from native speech recognition, nonnative speech recognition could be promoted by additional modeling using L2 speakers' speech. As a first step for the those task we analysed nonnative speech and then set a strategy for modeling Korean speakers'.

• Proceedings of the KSPS conference
• /
• 2007.05a
• /
• pp.327-329
• /
• 2007

This paper presents a FSN-based LVCSR system and it's application to the speech TV program guide. Unlike the most popular statistical language model-based system, we used FSN grammar based on the graph theory-based FSN optimization algorithm and knowledge-based advanced word boundary modeling. For the memory and latency efficiency, we implemented the dynamic pruning scheduling based on the histogram of active words and their likelihood distribution. We achieved a 10.7% word accuracy improvement with 57.3% speedup.

• Proceedings of the KSPS conference
• /
• 2007.05a
• /
• pp.330-333
• /
• 2007

This paper presents a dialogue interface using the dialogue management system as a method for controlling home appliances in Home Network Services. In order to realize this type of dialogue interface, we annotated 96,000 utterance pair sized dialogue set and developed an example-based dialogue system. This paper introduces the automatic error correction module for the SMS-styled sentence. With this module we increase the accuracy of NLU(Natural Language Understanding) module. Our NLU module shows an accuracy of 86.2%, which is an improvement of 5.25% over than the baseline. The task completeness of the proposed SMS dialogue interface was 82%.

• MALSORI
• /
• no.47
• /
• pp.131-140
• /
• 2003

We propose a new method of generating client models for HMM based text-dependent speaker verification system with only a small amount of training data. To make a client model, statistical methods such as segmental K-means algorithm are widely used, but they do not guarantee the quality or reliability of a model when only limited data are avaliable. In this paper, we propose a blind speech segmentation based on level building DTW algorithm as an alternative method to make a client model with limited data. In addition, considering the fact that voiced sounds have much more speaker-specific information than unvoiced sounds and energy of the former is higher than that of the latter, we also propose a new score evaluation method using the observation probability raised to the power of weighting factor estimated from the normalized log energy. Our experiment shows that the proposed methods are superior to conventional HMM based speaker verification system.