Improved scheme of run-graph encoding for efficient base presentation of line-drawing images
Zao Jiang, Jun'an Hu, Jiren Liu, Xiaoming Liu · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 1998
The run-graph is an efficient base representation of line drawing images, which is firstly proposed by Monagan and Roosli. It saves a large amount of memory space for storing a line drawing image whereas preserves all raster information of the original image, and the mapping of the image into a run-graph is bijective. It delivers the topological information of the image by constructing node area and edge areas. In this paper, we present an improved run-graph scheme with all efficiencies of the original run- graph preserved based on practices of run-graph representation. The improvements are mainly in two respects: (1) introducing an incline coefficient for judging the short run-length, which is equivalent to be capable to adjust the skew degree of the image, or change the position of a turning point. The characteristic of run-graph generation with the inclined coefficient is studied, and thus presents a simple method for deleting the redundancy turning points. (2) Presenting definitions of a protruding noise and a concave noise and their deleting methods. The experimental example confirms that the given scheme improves the accuracy of the mapping of the topological structure of a line drawing image into node and edge areas of run-graph representation. It provides an efficient information representation for furthermore processing and recognition.