期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2019
卷号：97
期号：19
页码：5070-5080
出版社：Journal of Theoretical and Applied
摘要：In many of the applications the content of an uneven illuminated images needs to be improved or recognized. For the degraded source images the global thresholding algorithm fails to produce adequate results. Due to this reason many applications used local thresholding techniques to binarize each pixel based on gray scale information of its neighborhood pixels. This paper discusses about the design and development of local thresholding techniques using specific fuzzy inclusion and entropy measures with fixed �r� and variable �r�. The noise influence on thresholding also tested using different noises like salt & pepper, Gaussian and speckle noises at different proportions. Different statistical parameters are evaluated to test the performance of the local thresholding algorithm with fixed �r� and variable �r�. It is evidenced from the results that the local thresholding method with variable �r� produced better results than compared to other methods.
关键词：Non-uniform Illumination; Global Thresholding; Local Thresholding; Fuzzy Inclusion; Entropy Measures
