Abstract
In this article, we evaluate the usefulness of using texture to segment a face image into Skin, Background and Hair areas. First, we present existing skin detection techniques, justify the a priori choice of texture, and then construct local feature indices from data learned from the image. We define a classification model based jointly on colour and texture data. We demonstrate the value of this approach by presenting some promising results obtained on complex natural images.
| Originalsprog | Fransk |
|---|---|
| Titel | ORASIS’09 Congrès des jeunes chercheurs en vision par ordinateur |
| Redaktører | Eric Marchand, Luce Morin, Alexandre Krupa, Edith Blin, Léna Baudoin, Claire Ammoniaux |
| Publikationsdato | 2009 |
| Status | Udgivet - 2009 |
Citationsformater
- APA
- Standard
- Harvard
- Vancouver
- Author
- BIBTEX
- RIS