Development of Shape-based 3D Object Automatic Annotation

2016 Fiscal Year Final Research Report

• Project/Area Number: 15K15992
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Multimedia database
• Research Institution: Toyohashi University of Technology
• Principal Investigator: Tatsuma Atsushi 豊橋技術科学大学, 工学(系)研究科(研究院), 助教 (60711166)
• Project Period (FY): 2015-04-01 – 2017-03-31
• Keywords: マルチメディア / パターン認識 / 情報検索 / 三次元物体

Outline of Final Research Achievements

We have witnessed a wide application of 3D objects (models) in various fields such as manufacturing, medicine, and entertainment. Accordingly, the development of 3D model retrieval techniques attracts attention. To realize a text-based search engine of 3D models, each 3D model needs to be given labels representing the shape of that. In this research, to develop shape-based 3D object automatic annotation technique, we have investigated as follows: (1) feature extraction method for 3D object based on deep learning and (2) dimensionality reduction method preserving relationships between 3D objects and labels evaluated with SimRank.

Free Research Field

マルチメディア情報検索