Çѱ¹Á¤º¸Åë½ÅÇÐȸ ³í¹®Áö (Journal of the Korea Institute of Information and Communication Engineering)
ÇѱÛÁ¦¸ñ(Korean Title) |
ÀâÀ½ÇÏÀÇ À½¼ºÀνÄÀ» À§ÇÑ ½ºÆåÆ®·² º¸»ó°ú ÁÖÆļö °¡Áß HMM |
¿µ¹®Á¦¸ñ(English Title) |
A Frequency Weighted HMM with Spectral Compensation for Noisy Speech Recognition |
ÀúÀÚ(Author) |
À̱¤¼®
Kwang-Seok Lee
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 05 NO. 03 PP. 0443 ~ 0449 (2001. 06) |
Çѱ۳»¿ë (Korean Abstract) |
ÀâÀ½È¯°æ¿¡¼ÀÇ À½¼ºÀνÄÀº ½ÇÁ¦ÀÇ È¯°æ¿¡¼ÀÇ À½¼ºÀνĿ¡¼ ¸Å¿ì Áß¿äÇÑ ¾Ö·Î±â¼ú·Î½á À̸¦ ÇØ°áÇϱâ À§ÇÑ ¿¬±¸´Â ²ÙÁØÈ÷ ¿¬±¸µÇ°í ÀÖ´Ù. µû¶ó¼ º» ¿¬±¸´Â À½¼ºÀνĺо߿¡¼ °¡Àå ¸¹ÀÌ »ç¿ëÇÏ°í ÀÖ´Â HMMó¸® ½ÃÀâÀ½Ã³¸®ÀÇ ¹®Á¦Á¡À» ÁÖÆļö °¡ÁßÄ¡ ºÎ°¡ HMMÀ¸·Î ÇØ°áÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÏ°í ±× ¼º´ÉÀ» ÀνĽÇÇèÀ» ÅëÇÏ¿© °ËÅäÇÏ¿´´Ù. ±× °á°ú SS󸮸¦ ÇÔ²² »ç¿ëÇÏ´Â MCE-u, MCE-p °¡ °¡Àå ÀâÀ½¿¡ °ÇÑ ¹æ½ÄÀÓÀ» ¾Ë ¼ö ÀÖ¾ú´Ù.
|
¿µ¹®³»¿ë (English Abstract) |
This paper is simulation research to improve speech recognition rates under the noisy environment. We examines recognition ratio based on frequency-weighted HMM together with spectral subtraction. As results, frequency-weighted HMM with scaling coefficients is trained as a minimum error classification criterion, and is presents a higher recognition rates in noisy condition than a conventional method. Furthermore, spectral subtraction method gives 11 to 28% improvements for this frequency-weighted HMM in low SNR, and gives recognition rates of 81.7% at 6dB SNR of noisy speech. |
Å°¿öµå(Keyword) |
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|