Face Detection using Skin-Color Filtering
| Autor: | Hsuan-Yu Wu, 吳宣諭 |
|---|---|
| Rok vydání: | 2006 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 94 It is a long history of human face detection and the technology is very maturation. However, it always used in security system such as face tracking and face recognition. The demands of images retrieve and multimedia entertainment has been increased very much due to the improvement of internet; therefore, it is foreseeable that the face detection field will have a big change when new applications arrives. There are four kinds of method in face detection field, which are knowledge-based, template-based, feature-based, and appearance-based. In this thesis, we proposed a method based on feature extraction and using skin color filtering to detect human faces in color images with complex background. We also try to realize the real-time ability that it may improve the performance of images retrieve and face tracking system. There are three steps in this system which are skin color detection, de-noise processing, and face verification. In skin color detection step, system inputs color images to filter out skin color region; in de-noise processing step, system preserve the most face-like region by morphology; after the first two steps, system outputs the face region images by using geometric estimation, feature extraction, and block correction. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|