Proposal of a Method to Extract Arbitrary FiguresUsing One-Dimensional Histograms
Shota Nakashima, Makoto Miyauchi, Seiichi Serikawa · Journal of Advanced Computational Intelligence and Intelligent Informatics · 2009
Arbitrary figure extraction, a basic image processing problem, is done typically using the generalized Hough transform (GHT). GHT and its successors tend, however, to consume humongous amounts of processing time and memory space. The arbitrary figure extraction we propose using one-dimensional histograms takes advantage of the Polytope method, which features: (1) The histogram distribution changes if parameters representing figures change. (2) Optimum parameters are obtained, if the value of the highest-frequency histogram becomes maximum. This approach makes memory space very small, processing time very short, effective by extracts arbitrary curves with different aspect ratios, and the algorithm is simple.