A Study on Geometry Contrast Enhancement for 3D Point Models

Abstract

Point primitives have come into the spotlight as a representation method of 3D models. A lot of researches have been performed on the modeling, processing, and rendering 3D point models. Especially, various methods have been developed for the extraction and preservation of the salient features of corners, curves, and edges in 3D point models. However, little effort has been made to extract and enhance the weak features that are relatively imperceptible due to the low geometry contrast. In this thesis, we propose a novel method to improve the visibility of 3D point models by enhancing the geometry contrast of weak features. We first define a weak feature region as a group of local points yielding small deviations of normal directions. Then we define the geometry histogram for each region as the distribution of the signed distance between a feature point and the locally approximated plane. We equalize and stretch the geometry histogram and move the corresponding feature points accordingly. We also render the enhanced model using the normal mapping for better visual presentation. Experimental results demonstrate that the proposed method enhances the geometry contrast of 3D point models by refining the appearance of the weak features. We expect that the geometry contrast enhancement algorithm will facilitate many applications in various fields.