Image precision silhouette edges

摘要: Finding and displaying silhouette edges is important in applications ranging from computer vision to nonphotorealistic rendering. To render visible of a polygonal object scene given viewpoint, we must first find all edges, i.e. boundaries between adjacent front facing back-facing surfaces. This followed by solving the partial visibility problem so that only those parts which are not occluded interior any surface, rendered. The may optionally be rendered with lighting model. paper describes simple general-purpose method combine three operations for composed objects can scan-converted. Using depth buffer, rendering process computes intersection surfaces image space at interactive rates. All performed image-precision hence special care taken limited numerical precision buffer. A solution suggested using view- dependent modification objects. does require preprocessing or adjacency information applicable dynamic scenes. Our based on observation two sets primitives needed compute second layer polygons viewpoint. In other words, these include just behind them either back facing. Intersection gives edges. this propose specifically models but it extended representation layers their computed. proposed nearest computed real time precision. robust, general purpose easy implement. No pre-processing required. allows use scenes, changing topology different levels detail traditional polygon pipeline. We also describe how used create interesting effects.