情報処理学会 第87回全国大会

2P-02
自然音の楽器性評価システム
○王 昱寰,伊藤克亘(法大)
This study proposes a system for evaluating the instrumentality of natural sounds using deep learning methods. By training neural networks to extract feature vectors of natural sounds and applying these to musical instrument datasets, we analyze the similarity between natural sounds and traditional instruments. The results demonstrate the potential of this method for distinguishing natural sounds and matching them with similar instrument types, providing new directions for sound design.