Understanding Human Activity in Kitchen via Food, Tool, and Action Recognition

2015 Fiscal Year Final Research Report

• Project/Area Number: 24240030
• Research Category: Grant-in-Aid for Scientific Research (A)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Perception information processing/Intelligent robotics
• Research Institution: Kyoto University
• Principal Investigator: Minoh Michihiko 京都大学, 学術情報メディアセンター, 教授 (70166099)
• Co-Investigator(Kenkyū-buntansha): MUKUNOKI Masayuki 京都大学, 学術情報メディアセンター, 准教授 (20283640) ; MORI Shinsuke 京都大学, 学術情報メディアセンター, 准教授 (90456773) ; YAMAKATA Yoko 京都大学, 情報学研究科, 准教授 (60423018) ; FUNATOMI Takuya 京都大学, 学術情報メディアセンター, 助教 (20452310) ; NAKAMURA Kazuaki 京都大学, 法学(政治学)研究科(研究院), 助手 (10584047) ; Atsushi Hashimoto 京都大学, 法学(政治学)研究科(研究院), 助手 (80641753)
• Co-Investigator(Renkei-kenkyūsha): NAKAMURA Kazuaki 大阪大学, 工学(系)研究科(研究院), 助教 (10584047) ; FUNATOMI Takuya 奈良先端科学技術大学院大学, 情報科学研究科, 准教授 (20452310)
• Project Period (FY): 2012-04-01 – 2016-03-31
• Keywords: 映像，文章，音声等認識 / 情報図書館学 / スマートセンサ情報システム / 知能ロボティクス / 行動認識 / 画像，文章，音声等認識

Outline of Final Research Achievements

This study aimed to understand human activity in kitchen in which various tools are used and decision-making is required; the activity represents human intelligence that separates us from other animals. In“visible actions” in observations, we focused on object grabbing and releasing actions, and developed a method to recognize completion of “invisible activities” that are linguistically described in recipe text and to forecast the forthcoming action intended by humans. We also developed several methods to extract “invisible activities” in recipe texts as workflow graph with its word class (i.e. food, tools, food-states and so on).

Free Research Field

知覚情報処理・知能ロボティクス