An Efficient JMC Algorithm for the Rhythm Query in Music Databases
碩士 === 國立中山大學 === 資訊工程學系研究所 === 97 === In recent years, the music has become more popular due to the evolution of the technology. Various kinds of music around us become more complexity and huge. This explosive growth in the music has generated the urgent need for new techniques and tools that can i...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2009
|
Online Access: | http://ndltd.ncl.edu.tw/handle/9736js |
id |
ndltd-TW-097NSYS5392017 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-097NSYS53920172019-05-29T03:42:53Z http://ndltd.ncl.edu.tw/handle/9736js An Efficient JMC Algorithm for the Rhythm Query in Music Databases 一個於音樂資料庫中有效率的JMC節奏查詢之方法 Han-ping Chou 周漢平 碩士 國立中山大學 資訊工程學系研究所 97 In recent years, the music has become more popular due to the evolution of the technology. Various kinds of music around us become more complexity and huge. This explosive growth in the music has generated the urgent need for new techniques and tools that can intelligently and automatically transform the music into useful information, and classify the music into correct music groups precisely. The rhythm query is the fundamental technique in music genre classification and content-based retrieval, which are crucial to multimedia applications. Recently, Christodoulakis et al. has proposed the CIRS algorithm that can be used to classify music duration sequences according to rhythms. In the CIRS algorithm, a rhythm is represented by a sequence of “Quick” (Q) and “Slow” (S) symbols, which corresponds to the (relative) duration of notes, such that S = 2Q. In order to classify music by rhythms, the CIRS algorithm locates the MaxCover which is the maximum-length substring of the music duration sequence, which can be covered (overlapping or consecutively) by the rhythm query continuously. During the matching step, one S symbol in the rhythm query can be regarded as two consecutive Q symbols in the duration sequence, but the two consecutive Q symbols in the rhythm query can not be combined as one S symbol in the duration sequence. This definition causes the difficulty for designing the algorithm. The CIRS algorithm contains four steps and repeat Steps 2, 3, and 4 to get local MaxCover for each different duration value of the music duration sequence. Finally, the global MaxCover is computed. We observe that it will generate unnecessary results repeatedly among Steps 2, 3, and 4. Therefore, in this thesis, to avoid repeatedly processing Steps 2, 3, and 4 for each different duration value, we propose the JMC (Jumping-by-MaxCover) algorithm which provides a pruning strategy to find the MaxCover incrementally, resulting in the reducing of the processing cost. In fact, we can make use of the relationship between the MaxCover MX founded by a different duration value X, and use the duration sequences cut by such a different duration value X to reduce the unnecessary process for the other different duration value Y , where Y < X. To make use of this property to reduce the processing time, we propose a cut-sequence structure and update it incrementally to compute the final global MaxCover. In this way, we can skip many steps and find the same answer of the CIRS algorithm. From our simulation results, we show that the running time of the JMC algorithm could be shorter than that of the CIRS algorithm. When the largest different duration value is uniformly distributed in the duration sequence, the running time can be reduced hugely, which is the best case of our proposed JMC algorithm. Ye-In Chang 張玉盈 2009 學位論文 ; thesis 88 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立中山大學 === 資訊工程學系研究所 === 97 === In recent years, the music has become more popular due to the evolution of the technology. Various kinds of music around us become more complexity and huge. This explosive growth in the music has generated the urgent need for new techniques and tools that can intelligently and automatically transform the music into useful information, and classify the music into correct music groups precisely. The rhythm query is the fundamental technique in music genre classification and content-based retrieval, which are crucial to multimedia applications. Recently, Christodoulakis et al. has proposed the CIRS algorithm that can be used to classify music duration sequences according to rhythms. In the CIRS algorithm, a rhythm is represented by a sequence of “Quick” (Q) and “Slow” (S) symbols, which corresponds to the (relative) duration of notes, such that S = 2Q. In order to classify music by rhythms, the CIRS algorithm locates the MaxCover which is the maximum-length substring of the music duration sequence, which can be covered (overlapping or consecutively) by the rhythm query continuously. During the matching step, one S symbol in the rhythm query can be regarded as two consecutive Q symbols in the duration sequence, but the two consecutive Q symbols in the rhythm query can not be combined as one S symbol in the duration sequence. This definition causes the difficulty for designing the algorithm. The CIRS algorithm contains four steps and repeat Steps 2, 3, and 4 to get local MaxCover for each different duration value of the music duration sequence. Finally, the global MaxCover is computed. We observe that it will generate unnecessary results repeatedly among Steps 2, 3, and 4. Therefore, in this thesis, to avoid repeatedly processing Steps 2, 3, and 4 for each different duration value, we propose the JMC (Jumping-by-MaxCover) algorithm which provides a pruning strategy to find the MaxCover incrementally, resulting in the reducing of the processing cost. In fact, we can make use of the relationship between the MaxCover MX founded by a different duration value X, and use the duration sequences cut by such a different duration value X to reduce the unnecessary process for the other different duration value Y , where Y < X. To make use of this property to reduce the processing time, we propose a cut-sequence structure and update it incrementally to compute the final global MaxCover. In this way, we can skip many steps and find the same answer of the CIRS algorithm. From our simulation results, we show that the running time of the JMC algorithm could be shorter than that of the CIRS algorithm. When the largest different duration value is uniformly distributed in the duration sequence, the running time can be reduced hugely, which is the best case of our proposed JMC algorithm.
|
author2 |
Ye-In Chang |
author_facet |
Ye-In Chang Han-ping Chou 周漢平 |
author |
Han-ping Chou 周漢平 |
spellingShingle |
Han-ping Chou 周漢平 An Efficient JMC Algorithm for the Rhythm Query in Music Databases |
author_sort |
Han-ping Chou |
title |
An Efficient JMC Algorithm for the Rhythm Query in Music Databases |
title_short |
An Efficient JMC Algorithm for the Rhythm Query in Music Databases |
title_full |
An Efficient JMC Algorithm for the Rhythm Query in Music Databases |
title_fullStr |
An Efficient JMC Algorithm for the Rhythm Query in Music Databases |
title_full_unstemmed |
An Efficient JMC Algorithm for the Rhythm Query in Music Databases |
title_sort |
efficient jmc algorithm for the rhythm query in music databases |
publishDate |
2009 |
url |
http://ndltd.ncl.edu.tw/handle/9736js |
work_keys_str_mv |
AT hanpingchou anefficientjmcalgorithmfortherhythmqueryinmusicdatabases AT zhōuhànpíng anefficientjmcalgorithmfortherhythmqueryinmusicdatabases AT hanpingchou yīgèyúyīnlèzīliàokùzhōngyǒuxiàolǜdejmcjiézòucháxúnzhīfāngfǎ AT zhōuhànpíng yīgèyúyīnlèzīliàokùzhōngyǒuxiàolǜdejmcjiézòucháxúnzhīfāngfǎ AT hanpingchou efficientjmcalgorithmfortherhythmqueryinmusicdatabases AT zhōuhànpíng efficientjmcalgorithmfortherhythmqueryinmusicdatabases |
_version_ |
1719192992760201216 |