• MALSORI
• /
• no.62
• /
• pp.97-112
• /
• 2007

The speech recognizer in general operates in noisy acoustical environments. Many research works have been done to cope with the acoustical variations. Among them, the multiple-HMM model approach seems to be quite effective compared with the conventional methods. In this paper, we consider a multiple-model approach combined with the model compensation method and investigate the necessary number of the HMM model sets through noisy speech recognition experiments. By using the data-driven Jacobian adaptation for the model compensation, the multiple-model approach with only a few model sets for each noise type could achieve comparable results with the re-training method.

• MALSORI
• /
• no.58
• /
• pp.45-66
• /
• 2006

In this paper, a rule-based (transformation-based) NE recognition system is proposed. This system uses Brill's rule inference approach. The performance of the rule-based system and IdentiFinder, one of most successful stochastic systems, are compared. In the baseline case (no punctuation and no capitalisation), both systems show almost equal performance. They also have similar performance in the case of additional information such as punctuation, capitalisation and name lists. The performances of both systems degrade linearly with the number of speech recognition errors, and their rates of degradation are almost equal. These results show that automatic rule inference is a viable alternative to the HMM-based approach to NE recognition, but it retains the advantages of a rule-based approach.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.119-122
• /
• 2003

In general, triangular shape filters are used in the filter bank when we get the MFCCs from the spectrum of speech signal. In [1], a new feature extraction approach is proposed, which uses specific filter shapes in the filter bank that are obtained from the spectrum of training speech data. In this approach, principal component analysis technique is applied to the spectrum of the training data to get the filter coefficients. In this paper, we carry out speech recognition experiments, using the new approach given in [1], for a large amount of telephone speech data, that is, the telephone speech database of Korean connected digit released by SITEC. Experimental results are discussed with our findings.

• Proceedings of the KSPS conference
• /
• 2004.05a
• /
• pp.91-97
• /
• 2004

We provide a brief overview of the area of speaker recognition, describing underlying techniques and current market review. We describe the techniques mainly based on GMM(gaussian mixture model) that is the most prevalent and effective approach. Following the technical overview, we will outline the market review of the area inside and outside of the country.

• MALSORI
• /
• no.48
• /
• pp.81-91
• /
• 2003

In this paper, we propose a speech recognition method based on independent component analysis (ICA) and spectral enhancement techniques. While ICA tris to separate speech signal from noisy speech using multiple channels, some noise remains by its algorithmic limitations. Spectral enhancement techniques can compensate for lack of ICA's signal separation ability. From the speech recognition experiments with instantaneous and convolved mixing environments, we show that the proposed approach gives much improved recognition accuracies than conventional methods.

• Proceedings of the KSPS conference
• /
• 2003.10a
• /
• pp.97-100
• /
• 2003

This paper addresses a method of convolutive source separation that based on SEONS (Second Order Nonstationary Source Separation) [1] that was originally developed for blind separation of instantaneous mixtures using nonstationarity. In order to tackle this problem, we transform the convolutive BSS problem into multiple short-term instantaneous problems in the frequency domain and separated the instantaneous mixtures in every frequency bin. Moreover, we also employ a H infinity filtering technique in order to reduce the sensor noise effect. Numerical experiments are provided to demonstrate the effectiveness of the proposed approach and compare its performances with existing methods.

• Proceedings of the KSPS conference
• /
• 1996.10a
• /
• pp.123-128
• /
• 1996

So far we have proposed the following constraint ranking for the (over-)application of the coda neutralization: (22) License family ≫ UE family ≫ IDENT-IO family ≫ Base-ID This analysis shows that only the surface level is enough to analyze the opaque behaviors of coda neutralization. Uniform Exponence constraint is worth further study since it can handle Consonant Cluster Simplification and underapplication of /t/-palatalization in Korean compounds in which morphemes before a stem are uniformly realized as one surface form: i.e., the output base form (S. Hong in preparation)(equation omitted)

• The Journal of Korean Academy of Prosthodontics
• /
• v.30 no.2
• /
• pp.191-202
• /
• 1992

The purpose of the present study is to inspect the satisfied degree of each oral function in geriatric patients with the shortened dental arch and when their prosthetic treatment is on schedule, provide some references to such treatment. For the approach to such study, 521 subjects were reviewed by grouping them according to the number of their remaining teeth, and masticatory function, phonetic function, facial change, and TMJ disorders were inspected and clarified through some questionnaires. Also through the questionnaires, the correlations between the geriatiric patients with the shortened dental arch and dentition and between the geriatiric patients with the Free-end RDP at the shortened dental arch and their oral function were found out with their satisfied degree of oral function. Results or findings from such study are as follows : 1. With regard to their satisfied degree of oral function, there was a significant difference of satisfaction between or among the group having only the anterior teeth and the group having the part of premolars and the group having even the part of molars, however no significant difference of satisfaction appeard between the group having 1st molars and the group having 2nd molars. 2. With regard to their satisfied degree of phonetic function, no significant difference appeared between or among the group having only the anterior part of teeth and the group having even the part of premolars and the group having even the part of molars, and with regard to their satisfied degree of facial change, no significant difference of satisfaction appeared between the group having the part of premolars and the group having even the part of molars. 3. With regard to their satisfied degree of masticatory function, phonetic function, TMJ disorders, and facial change, no significant difference appeared between the group having both the anterior part of teeth and the part of premolars and the group attached with the Free-end RPD on the same conditions of the afore-said group.

• Proceedings of the KSPS conference
• /
• 2000.07a
• /
• pp.353-363
• /
• 2000

The consonant system of English is relatively uniform throughout the English-speaking countries. Accents of English are mainly known to differ in terms of their vowel systems as well as in the phonetic realisations of vowel phonemes. The results of an acoustic study of vowel phonology of Japanese English, Singapore English and Indian English are presented, and an attempt is then made to compare the vowel phonology of these non-native varieties with that of Scottish English and RP. Various native varieties of English are thus shown to differ from each other in major ways, as much, perhaps, as the non-native varieties differ from the native varieties. Nevertheless, native speakers of English appear to be mutually intelligible to a degree that does not extend to non-native varieties. Obviously there are features that various native accents have in common which facilitate their mutual intelligibility, and these features are not shared by non-native accents. It is proposed that the foreign learner adopt certain core features of English in his pronunciation if he is to use English effectively as an international language. The common core that is significant in the communication process will be discussed. In conclusion, some pragmatic implications for the English language education in the new millennium will be articulated.

• MALSORI
• /
• no.57
• /
• pp.99-112
• /
• 2006

In this paper, we introduce two retrieval methods for photos with speech documents. We compare the pattern of speech query with those of speech documents recorded in digital cameras, and measure the similarities, and retrieve photos corresponding to the speech documents which have high similarity scores. As the first approach, a phoneme recognition scheme is used as the pre-processor for the pattern matching, and in the second one, the vector quantization (VQ) and the dynamic time warping (DTW) are applied to match the speech query with the documents in signal domain itself. Experimental results show that the performance of the first approach is highly dependent on that of phoneme recognition while the processing time is short. The second method provides a great improvement of performance. While the processing time is longer than that of the first method due to DTW, but we can reduce it by taking approximated methods.