3D Facial Image Registration by Using Geometric Feature Detection
碩士 === 義守大學 === 生物醫學工程學系碩士班 === 98 === Three-dimensional facial imaging is useful in surgical planning and evaluation. The images capture the geometry of faces such that the surgeons can visualize and measure the changes in facial shapes. In general, we can use a laser scanner to obtain the facial i...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2010
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/38426125323741916997 |
| Summary: | 碩士 === 義守大學 === 生物醫學工程學系碩士班 === 98 === Three-dimensional facial imaging is useful in surgical planning and evaluation. The images capture the geometry of faces such that the surgeons can visualize and measure the changes in facial shapes. In general, we can use a laser scanner to obtain the facial images. Since one laser scan only capture the face from one viewpoint, the surface invisible from the scanner is missing. Thereafter, we usually capture the face in several viewpoints and register the image from different angles. Large amount of vertices and the difference between viewpoints require sufficient computation time to accomplish the registration.
We use ICP algorithm as the basis to investigate how to accelerate three-dimensional registration of facial images. We develop a method to automatic selecting the facial feature and the experimental results show that reducing the samples of the model has significant influence the time of facial image registration.
|
|---|