• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë

Á¤º¸°úÇÐȸ ³í¹®Áö B : ¼ÒÇÁÆ®¿þ¾î ¹× ÀÀ¿ë

Current Result Document : 11 / 12

ÇѱÛÁ¦¸ñ(Korean Title) Çã¹Ö ÁúÀÇ ±â¹Ý À½¾Ç °Ë»öÀ» À§ÇÑ µ¿Àû Å¸ÀÓ ¿öÇÎÀÇ ¼º´É Çâ»ó ¹æ¹ý
¿µ¹®Á¦¸ñ(English Title) An Improved Dynamic Time Warping Method for Query by Humming
ÀúÀÚ(Author) ±èÀ±Áö   ¹ÚÁ¤Èñ   Yunji Kim   Cheong Hee Park  
¿ø¹®¼ö·Ïó(Citation) VOL 41 NO. 04 PP. 0318 ~ 0326 (2014. 04)
Çѱ۳»¿ë
(Korean Abstract)
Çã¹Ö ÁúÀÇ ±â¹Ý À½¾Ç °Ë»ö(Query by Singing/Humming, QBSH)´Â »ç¿ëÀÚ°¡ ³ë·¡³ª Çã¹ÖÀ¸·Î ÁúÀǸ¦ ÇÏ¿© À½¾Ç µ¥ÀÌÅͺ£À̽ºÀÇ ³ë·¡¸¦ °Ë»öÇÏ´Â ½Ã½ºÅÛÀÌ´Ù. Open-End Dynamic Time Warping (OEDTW)Àº QBSH ¿¬±¸¿¡ ÀϹÝÀûÀ¸·Î »ç¿ëµÇ´Â ¹æ¹ý ÁßÀÇ Çϳª·Î¼­, target songÀÇ ³¡À» °íÁ¤ÇÏÁö ¾ÊÀ½À¸·Î½á ÀÔ·ÂµÈ query¿Í °¡Àå Àß ¸ÅĪÀÌ µÇ´Â target songÀÇ ¾ÕºÎºÐÀ» Ž»öÇÏ´Â ¹æ¹ýÀÌ´Ù. ±×·¯³ª OEDTWÀÇ ¼öÇàÀ» ¿Ï·áÇϱâ Àü¿¡´Â queryÀÇ target song¿¡ ´ëÇÑ ÃÖÀûÀÇ Àü¿ªÀû ½ºÄÉÀÏÀ» °áÁ¤ÇÒ ¼ö ¾ø±â ¶§¹®¿¡ Á¶ º¯µ¿À» º¸¿ÏÇϱâ À§ÇÑ Æò±ÕÀ̵¿ÀÌ Á¦´ë·Î ÀÌ·ç¾îÁú ¼ö ¾ø´Ù´Â ´ÜÁ¡ÀÌ ÀÖ´Ù. º» ³í¹®¿¡¼­´Â Linear Scaling ¹æ¹ýÀ» ÀÌ¿ëÇÏ¿© ¾òÀ» ¼ö ÀÖ´Â ÃÖÀûÀÇ scaling factor¸¦ »ç¿ëÇÏ¿© OEDTWÀÇ ¼º´ÉÀ» Çâ»ó½ÃÅ°´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ¶ÇÇÑ, Á¦¾ÈµÈ ¹æ¹ýÀÇ ½Ã°£ º¹Àâµµ¸¦ ÁÙÀÏ ¼ö ÀÖµµ·Ï ¼±ÅÃÀû ¸ÅĪ Àü·«À» Àû¿ëÇÏ´Â ¹æ¹ýÀ» º¸ÀδÙ. Á¦¾ÈÇÑ ¹æ¹ýÀº ºÎÁ¤È®ÇÑ Çã¹Ö¿¡ ´ëÇؼ­ ºü¸£±â¿Í Á¶ º¯µ¿ÀÇ ¿µÇâÀ» ÃÖ¼ÒÈ­ÇÏ¿© ¸ÅĪÀ» Çϴµ¥ È¿°úÀûÀÌ°í, ¼±ÅÃÀû ¸ÅĪ Àü·«Àº ¸ÅĪ ½Ã°£À» ÇöÀúÈ÷ °¨¼Ò½ÃÅ°¸ç °Ë»ö Á¤È®µµ Ãø¸é¿¡¼­µµ ³ôÀº ¼º´ÉÀ» º¸ÀÓÀ» ½ÇÇèÀ» ÅëÇØ ÀÔÁõÇÑ´Ù.
¿µ¹®³»¿ë
(English Abstract)
Query by Singing/Humming(QBSH) is to retrieve songs in the music database using user¡¯s singing or humming. Open-End Dynamic Time Warping(OEDTW) is one of the methods which are commonly used in QBSH studies. By not fixing the end of target song in advance, OEDTW finds a part of a target song which is well matched with an input query. However, it has a limitation with mean shifting for overcoming key variation, because it cannot determine an optimal global scale factor before completing a matching process. In this paper, we propose a method for improving OEDTW performance using an optimal scaling factor taken from Linear Scaling phase. Also in order to reduce matching time, a strategy to select candidates is shown. Experimental results demonstrate that the proposed method is effective in minimizing the influence of tempo and key variation for inaccurate input queries and the selective matching strategy gives better search accuracy as well as reduced time complexity.
Å°¿öµå(Keyword) µ¿Àû ŸÀÓ ¿öÇΠ  ¼±Çü ½ºÄÉÀϸµ   Open-end µ¿Àû ŸÀÓ ¿öÇΠ  ³ë·¡/Çã¹Ö¿¡ ÀÇÇÑ ÁúÀÇ   dynamic time warping(DTW)   linear scaling(LS)   open-end dynamic time warping (OEDTW)   query by singing/humming(QBSH)  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå