A Study on Word Selection Method and Device Improvement for Improving Speech Recognition Rate of Speech-Language-impaired in Severe Noise Environment

심한 소음환경에서 언어장애인 음성 인식률 향상을 위한 단어선정 방법 및 장치 개선에 관한 연구

Abstract

Speech recognition rate is lowered even in a noisy environment, and it is difficult for a person with a speech disability or an inconvenient language to use it in a social life. In addition to improving the inconvenience of using the language, 280 words were selected using the word selection method which was improved when the word was selected considering the pronunciation characteristics of the language impaired. The MEMS development device used in the experiment was made considering material, lead wire type, length and direction. We improved the speech recognition rate by using the developed word selection method and the MEMS device developed to improve the speech recognition rate due to incorrect pronunciation and severe noise. The new method of selecting words and the mems device were improved and the results were included.

언어장애인, 언어 사용이 불편한 분들의 경우 조금의 잡음 환경에도 음성인식률이 저하되어 사회 생활시 어려움을 겪게 된다. 언어 사용 시 불편함을 장치로 개선시킴과 동시에, 언어 장애인의 발음 특성을 고려하여 단어 선정 시 자체 개선한 단어 선정 방법을 사용하여 280개 단어를 선정하였다. 실험에 사용된 MEMS 개발 장치는 재질, 유도선 종류, 길이, 방향을 고려하여 제작되었으며 잘못된 발음으로 인한 음성과 심한 소음에서 음성 인식률 향상을 위하여 개발된 MEMS 장치와 개발된 단어 선정 방법을 사용하여 개선시켰다. 개선 방법으론 새로운 단어 선정 방법과 mems 장치를 개선하여 진행하였으며 결과를 포함하였다.

Fig. 1 Consonant and vowel position structure

Fig. 2 Configuration step-by-step process

Fig. 3 Improved Variation observation method

Fig. 4 Improved word application example

Fig. 5 Step-specific process diagram

Fig. 6 Improved MEMS device 1

Fig. 7 Improved MEMS device 2

Fig. 8 Experiment environment

Fig. 9 Unmodified Waveform

Fig. 10 Transformation Waveform(metal+(×1) + vertical)

Fig. 11 Transformation Waveform(metal + (×1) + horizontal)

Fig. 12 Transformation Waveform(metal+(×3) + vertical)

Fig. 13 Transformation Waveform(metal + (×3) + horizontal)

Fig. 14 Transformation Waveform(non-metal + stringplace + (×1) + vertical)

Fig. 15 Transformation Waveform(non-metal + stringplace + (×1) + horizontal)

Fig. 16 Transformation Waveform(non-metal + stringplace + (×3) + vertical)

Fig. 17 Transformation Waveform(non-metal + stringplace + (×3) + horizontal)

Fig. 18 Transformation Waveform(non-metal + empty place+ (×1) + vertical)

Fig. 19 Transformation Waveform(on-metal + empty place+ (×1) + horizontal)

Fig. 20 Transformation Waveform(non-metal + emptyplace + (×3) + vertical)

Fig. 21 Transformation Waveform(non-metal + empty place + (×3) + horizontal)

Table. 1 Types and symptoms of language disorders

Table. 2 Decomposition/synthesis process (example)

Table. 3 Supposed Recognition experiment word

Table. 4 Word recognition number (consonant)

Table. 5 Word recognition number (vowel)

Table. 6 Recognition word(non-metal + empty place + (×1) + vertical)

Table. 7 Recognition word(non-metal + empty place + (×1) + horizontal)

Table. 8 Recognition word(non-metal + empty place + (×3) + vertical)

Table. 9 Recognition word(non-metal + empty place + (×3) + horizontal)

Table. 10 Recognition word(non-metal + string place + (×1) + vertical)

Table. 11 Recognition word(non-metal + string place + (×1) + horizontal)

Table. 12 Recognition word(non-metal + string place + (×3) + vertical)

Table. 13 Recognition word(non-metal + string place + (×3) + horizontal)

Table. 14 Recognition word(metal + (×1) + vertical)

Table. 15 Recognition word(metal + (×1) + horizontal)

Table. 16 Recognition word(metal + (×3) + vertical)

Table. 17 Recognition word(metal + (×3) + horizontal)

Table. 18 Unmodified