| Presentation | 2014-05-23 3D facial shape analysis based on statistical learning methods Misae NAKATSU, Xian-Hua HAN, Ryosuke KIMURA, Yen-Wei CHEN, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Recently, the relationship between gene and the facial morphology attracts substantial attention. A generic framework for analyzing facial morphology variation using scanned 3D landmarks was proposed in our previous work, which mainly includes three steps: registration and landmark correspondence, statistical analysis of 3D facial shape, ancestral classification using facial features. However, the used landmark corresponding method in the previous work was based on non-rigid transformation, which is very complicated and has high computational-cost, and the classification accuracy was less than 70%, which still has large space to be improved. This study proposes a simple and fast landmark corresponding strategy using cylindrical transformation; it easily adjusts the dimension of the shape representation vector. Furthermore, beside the shape variation features extracted by Principle Component Analysis (PCA), we propose a novel discriminated feature using mean hyperplane, and normalize the feature vector for reducing scale affect. Experiments on ancestral classification show that our proposed strategy can significantly improve the recognition performances compared to the conventional work. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Facial morphology / Gene / Statistical learning / Mean hyperplane / Normalization |
| Paper # | SIP2014-13,IE2014-13,PRMU2014-13,MI2014-13 |
| Date of Issue |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2014/5/15(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 | Pattern Recognition and Media Understanding (PRMU) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | 3D facial shape analysis based on statistical learning methods |
| Sub Title (in English) | |
| Keyword(1) | Facial morphology |
| Keyword(2) | Gene |
| Keyword(3) | Statistical learning |
| Keyword(4) | Mean hyperplane |
| Keyword(5) | Normalization |
| 1st Author's Name | Misae NAKATSU |
| 1st Author's Affiliation | Graduate school of Information Science and Engineering, Ritsumeikan University() |
| 2nd Author's Name | Xian-Hua HAN |
| 2nd Author's Affiliation | Graduate school of Information Science and Engineering, Ritsumeikan University |
| 3rd Author's Name | Ryosuke KIMURA |
| 3rd Author's Affiliation | Graduate School of Medicine, University of the Ryukyus |
| 4th Author's Name | Yen-Wei CHEN |
| 4th Author's Affiliation | Graduate school of Information Science and Engineering, Ritsumeikan University |
| Date | 2014-05-23 |
| Paper # | SIP2014-13,IE2014-13,PRMU2014-13,MI2014-13 |
| Volume (vol) | vol.114 |
| Number (no) | 41 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |