THRESHOLD BINARIZATION OF IMAGES BASED ON THE SKEWNESS AND KURTOSIS OF TRUNCATED DISTRIBUTIONS
I. S. Gruzman
Novosibirsk State Technical University, pr. Karla Marksa 20, Novosibirsk, 630092 Russia
isgruzman@gmail.com
Keywords: threshold, image binarization, skewness, kurtosis, truncated distribution
Subsection: ANALYSIS AND SYNTHESIS OF SIGNALS AND IMAGES
Abstract
An heuristic method for determining the binarization threshold is proposed which is stably operating for bimodal and unimodal histograms of observed images. A measure based on the skewness and kurtosis of truncated distributions is proposed to choose the threshold. Experimental studies have shown that binarization can be successfully performed regardless of differences in class size.
|
