Counterpoint Harmony Search: An Accurate Algorithm for the Blind Deconvolution of Binary Images
In this paper a novel algorithm, called Counterpoint Harmony Search (CHS), is presented for the simultaneous denoising and deconvolution of binary images. No prior information about the noise or blur shape and size is required, which makes so called blind deconvolution of binary images possible using CHS. CHS is based on the Harmony Search algorithm and inspired by the island model parallel genetic algorithm. We compare results from the CHS algorithm with recent results from modern blind binary image deconvolution algorithms. We achieve 100% accuracy in all our examples and show that CHS is robust even when deconvolving severely noisy images. These results improve significantly on results reported by a recent binary image deconvolution algorithm.
J. Fourie R. Green S. Mills
Computer Science and Software Engineering, University of Canterbury, Christchurch, New Zealand Areograph Limited,90 Crawford St Dunedin,New Zealand
国际会议
上海
英文
1117-1122
2010-10-20(万方平台首次上网日期,不代表论文的发表时间)