Title :
A new method for fast skeletonization of binary images on cellular processor arrays
Author :
Bin Wang ; Mroszczyk, Przemyslaw ; Dudek, Piotr
Author_Institution :
Sch. of Electr. & Electron. Eng., Univ. of Manchester, Manchester, UK
Abstract :
The proposed skeletonization algorithm, based on wave propagation principle, modifies the iterative thinning approach by replacing template matching with simple morphological operations, significantly decreasing the computational requirements. Conceptually, this method can be seen as binary wave-propagations in a 2-layer cellular network, where a trigger-wave propagating in one layer computes distances, and another layer is used to detect wave-collisions. Implementations of the algorithm on a fine-grain SIMD cellular processor array (SCAMP-5 vision chip), and on a dedicated VLSI asynchronous cellular processor array are presented. The experimental results conclude the paper.
Keywords :
VLSI; cellular arrays; image thinning; parallel processing; wave propagation; 2-layer cellular network; SCAMP-5 vision chip; VLSI asynchronous cellular processor array; binary image skeletonization algorithm; binary wave propagation principle; computational requirements; fine-grain SIMD cellular processor array; iterative thinning approach; morphological operations; template matching; trigger-wave propagation; wave-collision detection; Algorithm design and analysis; Equations; Hardware; Logic arrays; Logic gates; Parallel processing; Skeleton; CMOS; SIMD processor arrays; Skeletonization; asynchronous VLSI; vision chips;
Conference_Titel :
Cellular Nanoscale Networks and their Applications (CNNA), 2014 14th International Workshop on
Conference_Location :
Notre Dame, IN
DOI :
10.1109/CNNA.2014.6888649
