Studies On Medical Image Enhancement Algorithm Based On Fuzzy Theory

Medical imaging technology plays an increasingly important role in the development of modern medicine and clinical diagnosis and treatment on account of its economic,efficient,non-invasive characteristics and other advantages.However,due to different influence,the quality of images no matter obtained from what kind of medical imaging equipment are far from expectation,mainly with noise interference,unclear detail,low contrast and bad visual effect.These bad effects would lead to incorrect lesions location and diagnosis.Therefore,it is necessary to enhance the medical images to better its visual effects.Medical images feature inherent ambiguous information which cannot be described by accurate mathematic language,but a fuzzy system is capable of describing diverse and uncertain knowledge or information.Based on the analysis of existing fuzzy enhancement algorithms,improved algorithms were proposed in this paper to overcome the shortcomings.Experiments showed great superiority in detail maintenance and noise reduction.The main contents are as follows:1.The first part,we briefly introduce the status quo of some existing image enhancement algorithms both at home and abroad,then explain why it's important to introduce fuzzy theory into solving image enhancement problems.Finally,we analyze its further application in image enhancement.2.For the most part,image enhancement methods based on partial differential equations transform image into the gradient domain for processing.From gradient values,the characteristic of different pixel can be found and the diffusion coefficient can be determined adaptively as well.However,the gradient value is greatly affected by noise and also cannot effectively distinguish the smoothing area and detail region in an image.Aim at these weakness,and taking image fuzziness into consideration,an enhancement algorithm was proposed based on fuzzy entropy and variational partial differential equation.Firstly,the image is transformed into the fuzzy domain and then fuzzy domain gradient is defined instead of gray gradient.Secondly,we construct a fuzzy entropy which could reflect the spatial change rate of gray value,and use it to determine the diffusion coefficient of gradient.Finally,image is reconstructed in fuzzy domain using the variational method,and then converted into the gray domain to obtain the final enhanced image.Experiments are carried out on medical images.Results show that the improved algorithm can better distinguish the smoothing area and detail region in an image,meanwhile suppress the noise to some extent.3.An adaptive fuzzy contrast enhancement algorithm based on information measure is proposed.Analyzing the shortcomings of existing fuzzy contrast enhancement algorithms,this paper introduces the information measure to describe the characteristics of image edge.That distinguishes edge pixels and noise from their basic features,therefore the new algorithm has good anti-noise performance.The proposed method defines a contrast function based on the information measure and adaptively determines the magnification of the contrast according to the local image information measure,through which the membership values are modified.Then the membership values are transformed to gray level.Experiments show that the method is more sensitive to the edge and weak edge of the blurred image,and can effectively enhance the image detail and suppress the noise.