Font Size: a A A

Quantitative Analysis Of Glaucoma Lesions Based On OCT Images

Posted on:2016-04-13Degree:MasterType:Thesis
Country:ChinaCandidate:Y P WangFull Text:PDF
GTID:2134330461482841Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Glaucoma, a serious eye disease, a leading cause of blindness, can change the shape and state of optic disc and cup significantly. The measurement of cup-to-disc ratio is even more important in the detection of glaucoma. At present, more fashioned glaucoma detection methods are mostly based on colourful fundus images, however, spectral domain optical coherence tomography (SD-OCT) images are better in precision and reliability, the cup-to-disc ratio is more accurate subsequently. In SD-OCT images, the cup-to-disc ratio is more related to the breakpoints of retinal pigment epithelium (RPE), so the purpose of this study is to put forward a method to detect the RPE breakpoints and evaluate the cup-to-disc ratio in SD-OCT images.Firstly, the paper introduced advantages of SD-OCT images compared to others and clinical symptoms of glaucoma. Then we introduced the progress of glaucoma study based on fundus images or SD-OCT images and meanwhile, indicated the significance of our research.Then, the paper described related methods concretely, including preliminary and accurate detection, as well as some important steps. In the preliminary detection, image segmentation and flattening based on threshold and morphology were adopted, however, the precision of detected breakpoints was not satisfied. So. in accurate detection, other technologies were added, such as extracting examples, determining features, getting projection images and limiting target areas, as well as detecting RPE breakpoints using support vector machine(SVM) and extream learning machine(ELM) respectively, modifying breakpoints based on gradient of label matrix. Subsequently, RPE breakpoints were detected accurately and evaluated qualitatively.In the next step, some measurements and indexs indicating glaucoma were analyzed, such as the width of disc rim, the thickness of RNFL, cup-to-disc ratio and so on. Subsequently, we analyzed the rationality of using cup-to-disc ratio in evaluating glaucoma, calculated cup-to-disc ratio of our experimental data based on the breakpoints detected above.Lastly, we made a summary for the work have been done above, and introduced the advantages and disadvantages of our methods. Finally, we prospected the improvement work need to be done in future and gave some advice.
Keywords/Search Tags:glaucoma, spectral domain optical coherence tomography, cup-to-disc ratio, retinal pigment epithelium, support vector machine, extreme learning machine, breakpoint recognition
PDF Full Text Request
Related items