رسالة جامعية
Facial Model Extraction/Estimation for MPEG-4 Related Applications
العنوان: | Facial Model Extraction/Estimation for MPEG-4 Related Applications |
---|---|
العنوان البديل: | 適用於MPEG-4相關應用之臉部模型擷取及預測 |
المؤلفون: | Tsang G. Lin, 林倉亙 |
مرشدي الرسالة: | Chung J. Kuo, 郭鐘榮 |
سنة النشر: | 1999 |
المجموعة: | National Digital Library of Theses and Dissertations in Taiwan |
الوصف: | 87 Facial model coding is an integral part in MPEG-4 related applications. To generate facial model requires stereoscopic view of the face in the pre-processing stage. Although facial model can be successfully estimated from two stereo facial images, some estimated feature points experience large position error due to the occlusion effects. In this thesis, we propose two facial model estimation (FME) algorithms to update the facial model from a stereo image sequence such that correct feature point position can be obtained. Without the a priori information about the position of the face with respect to the camera, rotation axis and translation distance of the face movement, an accurate facial model can still be obtained by our schemes. In addition, our schemes do not require the precise camera parameters and avoid the tedious camera calibration. In other words, the facial model extraction/estimation is easy to implement. Furthermore the algorithms are adjusted for applications in single-view image sequence. This more general situation extends the utilization of the proposed algorithms in a variety of applications. The results convince us that the proposed facial model estimation algorithms are practical. |
Original Identifier: | 087CCU00442026 |
نوع الوثيقة: | 學位論文 ; thesis |
وصف الملف: | 52 |
الإتاحة: | http://ndltd.ncl.edu.tw/handle/40393279958107581387Test |
رقم الانضمام: | edsndl.TW.087CCU00442026 |
قاعدة البيانات: | Networked Digital Library of Theses & Dissertations |
الوصف غير متاح. |