Example-based NPR employing texture localization and self-similarity

2014 Fiscal Year Final Research Report

• Project/Area Number: 24500114
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Media informatics/Database
• Research Institution: Nagoya Institute of Technology
• Principal Investigator: NAKAMURA Tsuyoshi 名古屋工業大学, 工学(系)研究科(研究院), 准教授 (90303693)
• Co-Investigator(Renkei-kenkyūsha): HE Lifeng 愛知県立大学, 情報科学部, 教授 (80305515)
• Project Period (FY): 2012-04-01 – 2015-03-31
• Keywords: グラフィクス / 非写実的画像表現 / オノマトペ / コンピュータビジョン

Outline of Final Research Achievements

We have introduced texture localization and self-similarity to example-based non-photorealistic rendering. During three years we have been studying, we achieved some remarkable results. The first one is acceleration of texture synthesis which uses fast graph-cut algorithm. The algorithm can realize high-quality synthesis. The second one is interactive system hiring onomatopoeic voice inputs. The system provides users intuitive and simple operation to realize NPR. These results were presented in some workshops and published in a journal.

Free Research Field

ソフトコンピューティング