Research On The Key Technologies Of OCT Fundus Images Processing

Medicine is a kind of applied science closely related to human health,and in recent years medical imaging technology has become the fastest growing technology in the medical research field,and the accuracy of clinical diagnosis is increasingly enhanced.With the rapid development of computer science and information technology,medical imaging technology has made much progress.A series of advanced medical imaging equipment are available now,and the processing methods of medical images,such as de-noising,enhancement,segmentation have become current hot topics.This paper introduces the main methods of medical image processing,current research and applications,including medical image segmentation and its application,medical image registration and its application and medical image visualization technology.Then give a brief review of the history of medical image processing technology,and introduce classical medical image registration techniques and mature application case,explain the concept of medical image segmentation and introduce classical medical image segmentation techniques and mature application case,also simply introduce the three-dimensional medical image visualization technology.This paper analyzes the key and difficult problems in medical image processing,and on this basis,proposed these innovations:Optical coherence tomography(referred OCT)is a new optical imaging diagnostic techniques and it has developed rapidly in recent years,OCT fundus image get from measuring instrument can show the retinal function and structure information clearly.Medical studies have shown that the thickness of the retinal nerve fiber layer(RNFL)can be an important indicator of changes in macular degeneration,glaucoma and other neurodegenerative diseases in ophthalmology,and OCT measuring instrument is able to obtain the retinal nerve fiber layer thickness,as soon as possible to make glaucoma diagnosis.AO-OCT fundus image is the experiment data of this paper,and this paper analyzes the characteristics of OCT images to achieve image registration and segmentation.Use the microscopic space continuity of adjacent images to extract gray focus feature,and complete the image registration in the horizontal direction;then use the edge characteristics to complete image edge detection,and complete image registration in the vertical direction.This paper introduces OCT fundus image preprocessing techniques:median filtering,bilateral filtering and making original image gray.Propose a method combining bilateral filtering with median filtering method,using bilateral filter for filtering the low-frequency information,using median filtering for the high-frequency information,and introduce OCT fundus image layer segmentation process,level set method and active contour model(Snake model),then complete OCT fundus image layer segmentation after registration,and then analyze the results of segmentation.The RNFL thickness can be clearly seen from the results.Finally,according to the effect of OCT fundus image segmentation and registration,summarize OCT fundus image segmentation and registration methods,also summarize the advantages and disadvantages of the proposed algorithm,and discuss the development of medical image processing techniques.