Research on Analyzing Non-rigid Objects' Behaviors by Temporal Image Processing

2016 Fiscal Year Final Research Report

• Project/Area Number: 26330209
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Perceptual information processing
• Research Institution: Waseda University
• Principal Investigator: OHYA Jun 早稲田大学, 理工学術院, 教授 (90329152)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 動画像処理 / 動的動画像処理 / 認識 / 早期認識 / 剛体 / 手術 / 確率

Outline of Final Research Achievements

In recent years, towards the application to various applications, temporal image processing has been actively researched. This research focuses on the following two topics: (1) Extracting and tracking moving objects in dynamic, temporal images, and (2) Analyzing surgical situations based on temporal image processing. For (1), the following method is proposed: a moving RGBD camera acquiares temporal images that contain scenes in which rigid objects moves, and then each rigid object is extracted and tracked based on clustering the feature points. For (2), the following two methods are proposed: for recognizing surgeons' hand actions, a method based on Bag-of-words and learning; for early recogntion, a method, for which"sub-action" is introduced, based on SVM and probability computations. Experimental results show the validity of the above-mentioned proposed methods.

Free Research Field

画像情報処理