| Presentation | 2000/3/16 Vanishing Point Detection for Image Sequence using Feature Points Akihiro MINAGAWA, Yukihiko KOBAYASHI, Norio TAGAWA, Toshiyuki TANAKA, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | A method of vanishing point detection from sequential images is proposed. Conventional approaches for sequential images use methods of vanishing point detection for a single frame. However, in sequential images we can utilize inter-frame information, which may be useful in improving performance of the detection. The proposed method in this paper expends the previously proposed method for a single frame, which, based on a probability model, simultaneously estimates the vanishing point position and classifies feature points into line components supported by the vanishing point, in assuming and utilizing a priori knowledge for the vanishing point sequence in the images. The structure of the whole algorithm is a 2-level hierarchy of EM (Expectation Maximization) algorithms: In the upper level, maximum likelihood estimation of model parameters, including those for line components, and MAP (maximum a posteriori) estimation of the vanishing point position, as well as classification of feature points, are to be solved simultaneously via EM algorithm. In the lower lebel, EM algorithm is again used to solve the problem of MAP estimation and classification of feature points simultaneously. Computation is simplified by applying the saddle-point approximation in the E-step of the upper level EM algorithm. Effectiveness of the proposed method is confirmed by computer simulations. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | image sequence / vanishing point detection / ML(maximum likelihood)estimation / MAP estimation / EM alogorithm / saddle / point approximation |
| Paper # | NLC99-67,PRMU99-250 |
| Date of Issue |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2000/3/16(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 | Pattern Recognition and Media Understanding (PRMU) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Vanishing Point Detection for Image Sequence using Feature Points |
| Sub Title (in English) | |
| Keyword(1) | image sequence |
| Keyword(2) | vanishing point detection |
| Keyword(3) | ML(maximum likelihood)estimation |
| Keyword(4) | MAP estimation |
| Keyword(5) | EM alogorithm |
| Keyword(6) | saddle |
| Keyword(7) | point approximation |
| 1st Author's Name | Akihiro MINAGAWA |
| 1st Author's Affiliation | Graduate School of Engineering, Tokyo Metropolitan University() |
| 2nd Author's Name | Yukihiko KOBAYASHI |
| 2nd Author's Affiliation | Graduate School of Engineering, Tokyo Metropolitan University |
| 3rd Author's Name | Norio TAGAWA |
| 3rd Author's Affiliation | Graduate School of Engineering, Tokyo Metropolitan University |
| 4th Author's Name | Toshiyuki TANAKA |
| 4th Author's Affiliation | Graduate School of Engineering, Tokyo Metropolitan University |
| Date | 2000/3/16 |
| Paper # | NLC99-67,PRMU99-250 |
| Volume (vol) | vol.99 |
| Number (no) | 709 |
| Page | pp.pp.- |
| #Pages | 8 |
| Date of Issue | |