| Presentation | 2005-03-14 Geometric Inference from Images and Model Selection Kenichi KANATANI, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Model selection is to infer the law that underlies observed sensor data. Once the law is known, the next step is to optimally estimate the unknown parameters contained in the governing equation, which has been well studied by statisticians. In contrast, model selection is in a sense a transcendental problem : Although many principals have been proposed in the past, there is no logical ground to assert that one is better than others, and the choice is left to the user. The purpose of this lecture is to show how Akaike's AIC and Rissanen's MDL, the two selection criteria best known in statistics, can be applied to image processing and computer vision problems. First, it is pointed out that Akaike's AIC and Rissanen's MDL cannot be applied to geometric inference in their original forms. It is shown that if we go back to their starting principles and modify their derivations so that they fit into the framework of image processing and computer vision, we obtain the geometric AIC and the geometric MDL that the author proposed. In this lecture, we describe their essential distinction from Akaike's AIC and Rissanen's MDL in intuitive terms. Using easy-to-understand real image examples, we demonstrate how extensive their applications to geometric inference in image processing and computer vision problems are. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | factorization method / affine camera model / orthographic projection / weak perspective projection / paraperspective projection / AIC / MDL |
| Paper # | CAS2004-95,SIP2004-138,CS2004-231 |
| Date of Issue |
| Conference Information | |
| Committee | CS |
|---|---|
| Conference Date | 2005/3/7(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 | Communication Systems (CS) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Geometric Inference from Images and Model Selection |
| Sub Title (in English) | |
| Keyword(1) | factorization method |
| Keyword(2) | affine camera model |
| Keyword(3) | orthographic projection |
| Keyword(4) | weak perspective projection |
| Keyword(5) | paraperspective projection |
| Keyword(6) | AIC |
| Keyword(7) | MDL |
| 1st Author's Name | Kenichi KANATANI |
| 1st Author's Affiliation | Department of Information Technology, Okayama University() |
| Date | 2005-03-14 |
| Paper # | CAS2004-95,SIP2004-138,CS2004-231 |
| Volume (vol) | vol.104 |
| Number (no) | 720 |
| Page | pp.pp.- |
| #Pages | 5 |
| Date of Issue | |