Image Fuzzification And Segmentation By Intuitionistic Fuzzy Knowledge Measure

IFKM(intuitionistic fuzzy knowledge measure)Intuitionistic fuzzy knowledge measure plays an important role in uncertainty problems.Because of the structural characteristics of IFSs(intuitionistic fuzzy sets),knowledge measure can effectively make up for the lack of intuitionistic fuzzy entropy and reflect the fuzziness and hesitation in the process of studying uncertain problems.This paper aims to improve the ability of knowledge measure to solve complex problems.A new modeling method is proposed.This method improves the normalized Hamming-Hausdorff distance,and combines with TOPSIS(technique for order preference by similarity to ideal solution)to measure the amount of knowledge by the degree of closeness.The aim is to establish a unified axiomatic system and knowledge measure in accordance with the requirements of axiomatic system in different environments.The main research contents are as follows:(1)In this paper,an effective modeling method for knowledge measure is proposed.An IFKM model is constructed based on the improved HammingHausdorff distance and TOPSIS idea.Then the model is extended to IVIFSs(interval-valued IFSs)and degenerated back to the classical FSs(fuzzy sets).Based on the structural characteristics of intuitionistic fuzzy sets,this paper fully embodies the inner meaning of knowledge,unifies knowledge measure and its axiomatic definition in different environments,makes up for the inherent defects of intuitionistic fuzzy entropy,improves the structure system of knowledge measure,and explores the essence of knowledge measure.(2)The proposed model or method is applied to image segmentation and uncertain decision making to verify the feasibility and effectiveness of the proposed modeling method.According to the structure characteristics of IFS,a refined and efficient pixel classification rule and IFS algorithm are proposed,and then the proposed knowledge measure model is used to calculate the knowledge amount of image IFSs under different thresholds,and the optimal segmentation threshold is determined to achieve image segmentation.The above key results are extended to the IVIFSs environment and degenerated to the classical FSs environment.It provides a new research perspective and some different ideas for the practical application of knowledge measurement.The experimental results show that the performance of the knowledge-driven image threshold segmentation method in this paper is stable and reliable,and the generated binary graph has better performance indexes,which is obviously better than other similar algorithms.The effective results can also be obtained in the case of uncertain decision making.This paper unifies the existing forms of knowledge measure in different environments in order to better explore the nature of knowledge,and introduces the new theory of knowledge measure into the field of image processing for the first time,providing a new research perspective for the combination of image processing and knowledge amount.