Visually Salient Features Guided Mesh Meaningful Segmentation

During the last decade, digital geometry processing technologies of three-dimensional mesh model such as modeling,reconstruction and visualization have had great development. Faceing with advanced, high-precision scanning technology, the mesh model's complexity and data volumes are increasing rapidly. There are a variety of applications that benefit from breaking up a three-dimensional object into components. These applications include texture mapping, metamorphosis, three-dimensional shape retrieval, simplification and compression, and control skeleton extraction for key-frame animation. In this paper, we study around the three-dimensional mesh segmentation.First of all, a surrvey is given to summarize three-dimensional mesh model segmentation techniques, including latest achievements, classification, evaluation and application in this field. We also summarize a various of mesh segmentation methods.Secondly, we propose a meaningful mesh segmentation method which uses visually salient features to guide the mesh segmentation process. Since the features adopted are closely related to the psychology-based theory of visual salience, the decomposition process can appropriately mimic the function of a human visual system. After preprocessing step of mesh coarsening by edge collapse algorithm, with the utility that the whole protrusion part is represented by a few salient points, most works such as representitive points searching and core approximation can be done on the coarsed mesh. We only refine the final partitioning boundary on the original model, which resulting in a huge calculation speed. Experiments and results show that the algorithm can perform more efficiently and the segmentation is meaningful.