Study On Reversible Contrast Enhancement Techniques For Medical Images

The technique of contrast enhancement has been a hot topic in the field of image processing.Medical images play an important role in clinical diagnosis and medical analysis.However,the quality of medical images is often low due to the limitations such as equipment and other con-ditions,which is constraining the development and application of medical imaging technology.So,contrast enhancement of medical images has important significance for the use of medical images.Traditional methods of image contrast enhancement can achieve enhancement effects on image contrast,but most of the existing algorithms cause the loss of image information after contrast enhancement.The detailed information in medical images is often critical in the diagno-sis of a condition.It is important to enhance the image while keeping the detailed information intact.In this thesis,we will conduct the research work on reversible contrast enhancement algorithms for medical images.In the recent years,reversible image contrast enhancement methods have been proposed,which can be used to enhance image contrast without information loss.This advantage allows us to improve the quality of medical images while keeping all the information of the images to aid medical research and diagnosis.However,in the experiments,it was found that the existing methods have defects such as image distortion and poor enhancement effects,which eventually affect the image quality.In this thesis,we study the image distortion problem and improve the image quality of the enhanced results by improving the pre-processing methods.At the same time,the Grab Cut segmentation method is adopted to segment medical images.As a result,an effective interactive and reversible contrast enhancement scheme is generated for medical im-ages.We also introduce a neural network-based automatic segmentation algorithm to propose a reversible contrast enhancement algorithm for medical images in multiple brain tissues.In the comparative experiments on a set of chest X-ray images and a set of MR brain images,our method significantly outperforms the existing methods with respect to several evaluation met-rics.In summary,the reversible contrast enhancement method proposed in this thesis effectively improves the contrast and visual quality of the region of interest in medical images.