| Presentation | 2010-03-15 Gradient Features for Accuracy Improvement of Face Image Verification and Recognition Yushi KOMAKI, Wataru OHYAMA, Tetsushi WAKABAYASHI, Fumitaka KIMURA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We proposed a method that describes face appearance using gradient features for appearance based face identification using frontal face images. A person who was registered before is called "genuine" and a person who was not registered is called "impostor". Verification is a task which classifies a person in an input image as an insider or an outsider. And the person classified into insider is sent to the recognition process. The recognition is a task which recognizes who the person is. Either proposed gradient features or conventional grayscale features is extracted from frontal face image, the two methods are compared. In the experiment, we use frontal face images taken under three different situations: (1) different facial expression, (2) different illumination, (3) facial appearance change by aging. In verification and recognition, higher performatnce of gradient features is confirmed than grayscale features in all three situations. Equal Error Rate has been reduced from 27.88% to 15.93%. Recognition Error Rate has been reduced from 8.99% to 5.22%. The reason is that variation of gradient features by variation of facial location in an image, form change of facial parts, and global illumination change is small. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Biometrics / Face Image Identification / Verification / Recognition / Gradient Features |
| Paper # | PRMU2009-244,HIP2009-129 |
| Date of Issue |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2010/3/8(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) | Gradient Features for Accuracy Improvement of Face Image Verification and Recognition |
| Sub Title (in English) | |
| Keyword(1) | Biometrics |
| Keyword(2) | Face Image Identification |
| Keyword(3) | Verification |
| Keyword(4) | Recognition |
| Keyword(5) | Gradient Features |
| 1st Author's Name | Yushi KOMAKI |
| 1st Author's Affiliation | Graduate School of Engineering, Mie University() |
| 2nd Author's Name | Wataru OHYAMA |
| 2nd Author's Affiliation | Graduate School of Engineering, Mie University |
| 3rd Author's Name | Tetsushi WAKABAYASHI |
| 3rd Author's Affiliation | Graduate School of Engineering, Mie University |
| 4th Author's Name | Fumitaka KIMURA |
| 4th Author's Affiliation | Graduate School of Engineering, Mie University |
| Date | 2010-03-15 |
| Paper # | PRMU2009-244,HIP2009-129 |
| Volume (vol) | vol.109 |
| Number (no) | 470 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |