분할 기법을 이용한 3차원 얼굴 형상의 추출 = 3D Facial Reconstruction using Segmentation Method

다국어 초록 (Multilingual Abstract)

This paper proposes a 3D facial reconstruction method using segmentation method. 3D laser scanner, which is based on the laser triangulation, projects a sheet beam on the face in horizontal. At the same time, laser profile images are captured by CCD camera. Center profiles should be extracted from the all profile images. And the depths of those profiles are composed as a 3D face. Generally, when we use laser, it must be used with low wavelength to remove the harm to the body, especially eyes, It makes laser occluded at some region such as hair, chin and the edge of the face. It makes 3D reconstruction more difficult. Proposed method segments a 3D face into FSFS, SSFS and IFS depending on laser intensity reflection coefficient. Each segment was extracted with different algorithm. Finally, those are composed of 2D point cloud of facial depth. As a result, we can get a more reliable 3D face.