Title :
An Improved Run-Based Algorithm for Fast and Exact Boundary Extraction
Author :
Tao Hu ; Baoping Guo
Author_Institution :
Inst. of Optoelectron. Sci. & Eng., Huazhong Univ. of Sci. & Technol., Wuhan, China
Abstract :
Capson proposed a run-based boundary extraction algorithm which can extract boundaries in one raster scan, the algorithm is very suitable for pipeline and parallel processing and can achieve high performance, but it has two problems which limit its application, one is the distortion at concave potion of the boundaries, and another is that it dose not address how to produce the detailed boundaries. This paper intends to solve the two problems respectively by introducing splitting point and merging point and by skillfully designing an extended data structure to produce both the 4-connected and 8-connected detailed boundaries expediently, and an improved algorithm is proposed to implement the fast and exact boundaries extraction. Experimental results show that the improved algorithm runs faster in comparison with the other two algorithms in ref..
Keywords :
data structures; edge detection; parallel algorithms; pipeline processing; data structure; image processing; parallel processing; pattern recognition; pipeline processing; run-based boundary extraction algorithm; Algorithm design and analysis; Application software; Data mining; Data structures; Image coding; Information systems; Merging; Parallel processing; Pattern recognition; Pipelines; RLE; boundary extraction; merging point; pattern recognition; splitting point; turning-points;
Conference_Titel :
Industrial and Information Systems, 2009. IIS '09. International Conference on
Conference_Location :
Haikou
Print_ISBN :
978-0-7695-3618-7
DOI :
10.1109/IIS.2009.73