Presentation | 2008/3/15 Object Concept Modeling Based on Relationship among Visual Information, Usage and Function Yasuhito SHINCHI, Takayuki NAGAI, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | A novel object concept model, which encodes relationship among visual information, function and usage, is proposed in this report. Essential attribute of an object (tool) is its function that achieves a particular purpose. Therefore, a function model is first constructed through observations from a camera. Function is defined as changes in a target object before and after the tool use. At the same time, the usage model is constructed from observations of hand shape, grasping arts and contact points of the tool. And then, the proposed system learns object concept that is based on the relationship among visual information, and learned function and usage models. Visual information is represented by SIFT(Scale Invariant Feature Transform). Since the proposed models are based on graphical model, it is possible for the system to infer unobservable information using observed one stochastically. For example, the system can infer the usage and/or function of the tool visually through the proposed model. Some experimental results using a system, in which the proposed model is implemented, are shown to validate the proposed model. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Object recognition / graphical model / variational Bayes / EM algorithm |
Paper # | HIP2007-159 |
Date of Issue |
Conference Information | |
Committee | HIP |
---|---|
Conference Date | 2008/3/15(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | |
Vice Chair | |
Secretary | |
Assistant |
Paper Information | |
Registration To | Human Information Processing (HIP) |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Object Concept Modeling Based on Relationship among Visual Information, Usage and Function |
Sub Title (in English) | |
Keyword(1) | Object recognition |
Keyword(2) | graphical model |
Keyword(3) | variational Bayes |
Keyword(4) | EM algorithm |
1st Author's Name | Yasuhito SHINCHI |
1st Author's Affiliation | Dept. of Electronic Engineering, The University of Electro-Communications() |
2nd Author's Name | Takayuki NAGAI |
2nd Author's Affiliation | Dept. of Electronic Engineering, The University of Electro-Communications |
Date | 2008/3/15 |
Paper # | HIP2007-159 |
Volume (vol) | vol.107 |
Number (no) | 553 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |