Automated Segmentation Of The Retinal Layers In Optical Coherence Tomography Images Using Continuous Maximum Flow Algorithm

With the growing of incidence in eye diseases, the diagnosis and research of eye diseases has been extremely widespread concerned. And it is the critical step in ophthalmology medical aided diagnosis system to segment the retinal layers in fundus images.Medical imaging plays an increasingly important role in the clinical diagnosis and treatment in the developing of modern medical technology, which provides an important scientific basis for diagnosis and treatment of disease. Optical coherence tomography(OCT) obtains widely attention and application in the ophthalmology, as a new medical imaging technology, it has the advantages of non-contact, non-invasive, real-time in-vivo imaging and so on. It could obtain clear image of retinal structure, monitor the subtle changes in the retina. So, it could detect a variety of diseases through analysis of the abnormal structure information about retinal thickness, shape, etc.Traditional retinal layers segmentation in fundus OCT image is mainly based on manual segmentation, which is time-consuming, labor-intensive, low efficiency. Furthermore, the results are largely depended on the clinician's prior knowledge, work status, experience and other subjective factors, resulting in the poor repeatability. According to the current medical situation, there are a large number of medical image data need to be processed in time. However, the doctor resources are insufficient, manual segmentation, which has been unable to meet the needs of modern clinical medicine. Therefore, computer-aided medical diagnosis and fundus OCT image retinal layer automatic segmentation has become a hot research topic at home and abroad. It is helpful to improve the efficiency of medical image processing, segmentation results, the accuracy of diagnosis and treatment of disease, and reduce the pressure on the doctor's work.In the structure characteristics of retina in fundus OCT images, this paper mainly studies the automatic segmentation algorithm for retinal layers in OCT fundus image based on continuous maximum flow algorithm.Firstly, the paper briefly introduces the background and purpose of retinal layers segmentation method in the fundus OCT image, and describes and analyzes current the domestic and foreign research conditions; Secondly, this paper introduces the knowledge of OCT technology and fundus image; And then we analyze the continuous maximum flow algorithm, we present a improved method based on continuous maximum flow algorithm to segment retinal layers automatically in fundus OCT image. On the basis of the original image gray processing to flatten the image, we get smooth and flat retinal image for reducing the memory and time required for segmentation and improving the efficiency and accuracy. Then we use the improved continuous maximum flow algorithm to segment image coarsely, obtaining seven retinal layer boundaries except INL-OPL and OS-RPE, then extract the two boundary area respectively for segmenting them separately, and the final result could be obtained by superimposing all retinal layer boundaries. Finally, the comparison and analysis of experimental results show the proposed method could segment all retinal layers in the OCT fundus image with high accuracy and reliability.