Research And Application On Segmentation Algorithm Of OCT Medical Image

Optical Coherence Tomography(OCT)is a high-resolution,non-contact,noninvasive biological tissue imaging technique.It could perform tomography on the organization to clearly distinguish the organizational structures.Its principle is that different organizations have different ability to reflect,absorb and scatter the light.OCT plays a similar role in the pathological observation of living tissues.Therefore,it is more and more widely used in clinical medicine.In order to make the OCT to be a more practical medical diagnosis system and improve the diagnostic efficiency of doctors,first,we did some research on the OCT software system,and we had implemented the OCT image acquisition,storage and display.On this basis,the OCT image would be further processed and analyzed.The related image segmentation algorithm is used for implementing the automatic segmentation of OCT images in this paper.The main work is as follows:We studied the optical coherence tomography system for respiratory detection.This system have many functions,such as OCT image acquisition,branch fiber video capture,scanning control,case information management,image display and processing.Then,we designed and implemented the functions of image display and processing.In this paper,we focused on the functions about original document analysis,image data coordinate conversion and pseudo-color enhancement.And other functions including image contrast,image scaling,image smoothing and so on were developed on this basis.Doctors usually need to measure the length,area,thickness and other parameters of the organizations.Through these parameters,doctors could know lesions of the lesions.Comparison of the parameters before and after treatment could assess the quality of treatment.For the purpose of improving the diagnostic efficiency,we combined with Otsu algorithm,mathematical morphology,edge detection and other methods to achieve the cavity of the wall of the automatic extraction and measurement according to the characteristics of OCT images.In addition,for the specific demands of the system,we used C# and C++ mixed programming way to develop the OCT image measurement software.Research on image segmentation based on level set method.We focused on the C-V model and Li Chunming model in the level set method,and combined with two models to achieve the OCT image of the automatic segmentation method.The experimental results showed that this method can be applied to the automatic segmentation of OCT images.Compared with other medical image segmentation algorithms,this method has stronger performance to degrade noise and better segmentation result.