OS12-1 Restoration of Guqin Music by Deep Learning Methods

Title:	OS12-1 Restoration of Guqin Music by Deep Learning Methods
Publication:	ICAROB2024
Volume:	29
Pages:	320-324
ISSN:	2188-7829
DOI:	10.5954/ICAROB.2024.OS12-1
Author(s):	Takashi Kuremoto, Kazuma Fujino, Hirokazu Takahashi, Shun Kuremoto, Mamiko Koshiba, Hiroo Hieda, Shingo Mabu
Publication Date:	February 22, 2024
Keywords:	deep learning, VGG16, YOLOv5, SVM, Guqin, Jianzi Pu, AI music
Abstract:	Guqin （古琴）music played an important role in the history of Asia cultures. The notation of Guqin ancient music remained more than 600, however, only about 100 of them are played in nowadays. The reason is that the handwritten Guqin notations named "Jianzi Pu" is hard to be understood, however, we challenge to restore the Guqin music by deep learning methods and few Jianzi Pu images. VGG16 and YOLOv5 were adopted in the recognition experiment for Guqin music restoration. For a well-known Guqin music "Sen-O-So"（仙翁操）, 55 kinds of single characters of Jianzi Pu and 4,951 images of them were collected from 23 kinds of Sen-O-So version found by the Internet or obtained by image processing such as rotation, enlarge (zoom-in), reduce (zoom-out), filtering, etc. The average accuracies of VGG16 and YOLOv5 were 87.50% and 88.47% for the test data, respectively. Additionally, it was realized an online output of Guqin music as its output of audio or video forms by YOLOv5 in this study.
PDF File:	https://alife-robotics.co.jp/members2024/icarob/data/html/data/OS/OS12-1.pdf
Copyright:	© The authors. This article is distributed under the terms of the Creative Commons Attribution License 4.0, which permits non-commercial use, distribution and reproduction in any medium, provided the original work is properly cited. See for details: https://creativecommons.org/licenses/by-nc/4.0/

HOME