Histogram method of image binarization based on fuzzy pixel representation
Year: 2017
Type: conference-paper
Abstract: Binarization is an important task in image processing. Numerous approaches to this problem are known: the thresholding algorithm, the Otsu method, etc. A promising direction in image processing is the use of fuzzy logic methods and the theory of fuzzy sets. Their use makes it possible to improve the quality of processing by providing information in a fuzzy form. Most of the existing fuzzy image processing techniques perform a transfer to fuzzy sets in the late stages of processing and analysis, which means some loss of useful information. In this paper a new method of binarization using fuzzy representation of an image based on fuzzy pixels is proposed. In this case, the original image is translated into a fuzzy view on the first step. Various methods for implementing this transformation are described. The membership functions for describing fuzzy pixels are discussed, and the requirements for their form and appearance are given. A description of the new binarization method using histograms and a developed fuzzy representation of the image is presented. The experimental verification of the developed method on a test example is carried out. The results were compared with the Otsu method. Developed algorithm showed stable work on noisy and ordinary pictures. To improve the quality in the future, more complex 5×5 and 7×7 masks can be used to obtain fuzzy pixels. The investigated approach showed its applicability in image processing problems. In the future, the fuzzy model of the image can be expanded to use fuzzy features of the second and higher types.
Authors: Egor Pugin, Arkady Zhiznyakov
Institution: Vladimir State University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W2777077293