Skip to Main Content
This paper applies a mesh segmentation algorithm to extract the region of interest (RoI) of a 3D model. While studied in image and video processing, RoI is a totally new research area in geometry processing. Mesh segmentation is one of the most important area in 3D geometry processing and also works as the pre-process to several other applications in computer graphics. In this paper, we are trying to utilize the mesh segmentation algorithm to realize the RoI effects in 3D Model rendering. The proposed algorithm extends the current curvature-based method and integrates the hierarchical clustering. Firstly, our algorithm extracts the geometric features of 3D models and take advantage of these geometrical hints to cluster the triangle faces by Fitting pertinent primitives. Lastly, we evaluate the centeroid of each region of 3D model that is about to be the focus point in RoI rendering which is realized by GPU-based depth of field technique.
Date of Conference: 23-25 Nov. 2010