透過您的圖書館登入
IP:216.73.216.134
  • 期刊

最佳化梅爾倒頻譜係數之研究及其於音樂曲風辨識之應用

On the Optimal MFCC and Its Applications to the Music Genre Classification

摘要


在音樂辨識系統中,梅爾倒頻譜係數被廣泛的運用在音訊及音樂資料庫的分類決策,而隨著梅爾倒頻譜係數的運用,不少梅爾倒頻譜係數的改良方法也隨之被提出,其中針對三角濾波器能量組進行權重調整以提升辨識效果的方法被證實是可行的。本篇論文嘗試以基因演算法尋找最佳辨識效果之權重,並應用在曲風辨識上。實驗過程針對比賽用之資料庫ISMIR,並得到明顯的改良效果。

並列摘要


Due to the wide applications of MFCCs in the audio signal processing, a lot of studies on the improvement of MFCCs were presented. Among them, the improvement of the weightings of the triangular filters is a worth study. In this study, genetic algorithms were applied to search the optimal weights and then applied to the music genre classification. Experiment results with the ISMIR music database show the improvement of the proposed method.

被引用紀錄


陳秀慧(2013)。應用基因演算法於RFM模型權重最佳化之研究〔碩士論文,國立臺北科技大學〕。華藝線上圖書館。https://doi.org/10.6841/NTUT.2013.00760
郭又禎(2014)。改良式梅爾倒頻譜參數應用於關鍵字萃取〔碩士論文,國立中央大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0031-0412201511591189

延伸閱讀