Proposal of Representation Method for Time-Series Media Based on Tree Structure and Realization of its Calculus

2016 Fiscal Year Final Research Report

• Project/Area Number: 26280089
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Intelligent informatics
• Research Institution: Future University-Hakodate
• Principal Investigator: HIRATA Keiji 公立はこだて未来大学, システム情報科学部, 教授 (30396121)
• Co-Investigator(Kenkyū-buntansha): 東条 敏 北陸先端科学技術大学院大学, 情報科学研究科, 教授 (90272989) ; 浜中 雅俊 京都大学, 医学(系)研究科(研究院), 研究員 (30451686) ; 長尾 確 名古屋大学, 情報科学研究科, 教授 (70343209) ; 北原 鉄朗 日本大学, 文理学部, 准教授 (00454710) ; 松原 正樹 筑波大学, 図書館情報メディア研究科(系), 特任助教 (90714494) ; 吉井 和佳 京都大学, 情報学研究科, 講師 (20510001)
• Co-Investigator(Renkei-kenkyūsha): TERASAWA Hiroko 筑波大学, 図書館情報メディア研究科(系), 助教 (70579094)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 知能情報処理 / 時系列メディア

Outline of Final Research Achievements

We have been pursuing the formalization and construction of an algebraic system for a time-span tree provided by music theory GTTM as a typical example for representing intention. We have clarified the theoretical properties of a musical morphing algorithm, and proposed an approximating solution of a relative suedo complement using the flip operation newly introduced. As for musical applications employing tree structures, we have prototyped a composition system using melodic contour and a arrangement system reducing the voices included in polyphony. As for a musical analyizer, we have invented an new algorithm adopting a deep learning, and achieved a high precision rate of 80% to 90%. As for the deployment towards different media, we have proposed a discussion time-span tree and implemented an application system.

Free Research Field

情報学