| 講演抄録/キーワード |
| 講演名 |
2022-12-21 10:40
Density-based Bias-Free Automatic Chart Generation for Rhythm Games ○Zhao Yifan・Tsunenori Mine(Kyushu Univ.) AI2022-35 |
| 抄録 |
(和) |
(まだ登録されていません) |
| (英) |
Automatic chart generation methods for rhythm games often use both acoustic features and difficulty information when building models to predict the onset of the chart. Since difficulty information relates to both onset placement and the key structure of the game, direct use of difficulty information may cause unnecessary information bias in onset prediction. In this paper, we propose an approach that calculates density information to generate charts of different difficulty levels and uses the calculated information as an attribute for each level of onset. This allows the model to learn onset distributions with different levels of difficulty without directly using the difficulty information. Furthermore, integrating density-based onset sequences into a single one improves prediction performance, and density information can also be used to filter charts of different difficulty levels from the integrated chart. |
| キーワード |
(和) |
/ / / / / / / |
| (英) |
Automatic Content Generation / Video Game / Machine Learning / / / / / |
| 文献情報 |
信学技報, vol. 122, no. 322, AI2022-35, pp. 13-17, 2022年12月. |
| 資料番号 |
AI2022-35 |
| 発行日 |
2022-12-14 (AI) |
| ISSN |
Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| PDFダウンロード |
AI2022-35 |