Reduction of false microaneurysms in retinal fundus images using fuzzy C means clustering in terms NLM anisotropic filter

Abstract

This thesis deals with the automatic Detection of micro aneurysms (MAs) in colour fundus images and its application to the diagnosis of Diabetic Retinopathy. Diabetic Retinopathy - a complication of diabetes mellitus - is a severe and wide-spread eye disease: it is the leading cause of blindness for the working age population in western countries. For the diagnosis of Diabetic Retinopathy, digital colour fundus images are becoming increasingly important. This fact opens the possibility of applying image processing techniques in order to facilitate and improve diagnosis in different ways. First, manual analysis can be improved by using image enhancement methods. Second, the main problem of diagnosing Diabetic Retinopathy is the identification of micro aneurysms in colour fundus images. Third, the process of extracting micro-aneurysms from non-micro-aneurysms candidates. This comparison is essential for a monitoring of the disease and it is done by FCM technique. FCM-NLM anisotropic based diagnosis improvement has been studied in this thesis, and algorithm within this framework has been developed. This work presents image enhancement methods like contrast enhancement and histogram processing.A method for the automatic detection of micro aneurysms (MAs) in colour retinal images is proposed in this paper. The recognition of MAs is an essential step in the diagnosis and grading of diabetic retinopathy. The proposed method realizes MAs detection through the fuzzy c means clustering strategy and NLM anisotropic filter .The final MA candidates can be extracted from the clustered data by connected components method. Testing on various simulated retina data repositories, the developed method presents promising results and show robustness to rotations and scale changes. Compared to Rotational cross sectional analysis for detection of micro aneurysms, the proposed method is getting the ROC score of 0.435. The proposed method has been tested in the Retinopathy Online Challenge, where it is proved to be competitive with the state-of-the-art approaches. With the algorithms developed in this thesis, it is possible to identify the false micro aneurysms in retinal fundus images that leads to conceive diagnostic tools that may play a major role very effectively monitoring of Diabetic Retinopathy. Even in countries that show a considerable lack of specialists in ophthalmology. newline