Efficient and Automated Detection of Microaneurysms from Non-Dilated Fundus Images
1 College of Mechanical & Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, 210016, China
2 Department of Ophthalmology, Jiangsu province hospital of TCM, Nanjing, 210029, China
Abstract：In order to automatically detect microaneurysms from non-dilated fundus images, and develop an automated diabetic retinopathy screening system, a novel, simple and efficient algorithm of automatic microaneurysms detection was established and studied in this work. Hard exudates and vessel were segmented by mathematical morphology on the basis of preprocessed green channel of the original non-dilated fundus image in RGB channels. Then, candidate regions of microaneurysms were obtained by removing hard exudates and vessel from the resulting image of extended-minima transform on the previously preprocessed image. Thus, the true lesions of microaneurysms were separated based on size information. The algorithm was tested on two groups of non-dilated fundus images with different quality. Statistical analysis of detection results showed that precision for the two groups were both high, each relative error of corresponding indexes between two groups was lower than 4%, and processing efficiency was high for which the mean time cost for processing an image is 9.7 seconds. Results suggest that the algorithm can efficiently detect microaneurysms from non-dilated fundus images, and it is stable and reliable. As a result, the proposed algorithm has high practical value.
［1］Wild S, Roglic G, Green A, et al. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030［J］. Diabetes Care, 2004, 27:1047-1053.
［2］Watkins PJ. ABC of diabetic retinopathy［J］. British Medical Journal, 2003, 7(2):105-107.
［3］Villarroel M, Ciudin A, Hernndez C, et al. Neurodegeneration: An early event of diabetic retinopathy［J］. World J Diabetes, 2010, 1(2): 57-64.
［4］彭金娟, 邹海东, 王伟伟, 等. 上海市北新泾社区糖尿病视网膜病变远程筛查系统的应用研究［J］. 中华眼科杂志, 2010, 46(3): 258-262.
［5］蒋升, 张莉, 克麦尔〖DK〗·艾则孜. 免散瞳眼底照相技术在糖尿病视网膜病变筛查中的应用研究［J］. 国际眼科杂志, 2008, 8(10): 2037-2039.
［6］陈喆, 张士胜, 朱惠敏. 糖尿病视网膜病变的国际临床分类分析［J］. 国际眼科杂志, 2011, 11(8): 1394-1401.
［7］Dupas B, Walter T, Erginay A, et al. Evaluation of automated fundus photograph analysis algorithms for detecting microaneurysms, haemorrhages and exudates, and of a computerassisted diagnostic system for grading diabetic retinopathy［J］. Diabetes & Metabolism, 2010, 36(3): 213-220.
［8］Spencer T, Olson JA, McHardy KC, et al. An imageprocessing strategy for the segmentation and quantification of microaneurysms in fluorescein angiograms of the ocular fundus［J］. Comp Biomed Res, 1996, 29: 284-302.
［9］Cree MJ, Olson JA, McHardy KC, et al. A fully automated comparative microaneurysm digital detection system［J］. Eye, 1997, 11: 622-628.
［10］Frame A, Undrill P, Cree M, et al. A comparison of computer based classification methods applied to the detection of microaneurysms in ophthalmic fluorescein angiograms［J］. Comput Biol Med,1998, 28:225-238.
［11］Hipwell JH, Strachan F, Olson JA, et al. Automated detection of microaneurysms in digital redfree photographs: a diabetic retinopathy screening tool［J］. Diabetic Medicine, 2000, 17:588-594.
［12］Sinthanayothin C, Boyce JF, Williamson TH, et al. Automated detection of diabetic retinopathy on digital fundus image［J］. Diabetic Medicine, 2002,19(2): 105-112.
［13］Larsen M, Godt J, Larsen N, et al. Automated detection of fundus photographic red lesions in diabetic retinopathy［J］. Investigat Ophthalmol Vis Sci, 2003, 44(2): 761-766.
［14］Niemeijer M, Ginneken B, Staal J, et al. Automatic detection of red lesions in digital color fundus photographs［J］. IEEE Trans Med Imaging, 2005, 24(5): 584-592.
［15］Martins CIO, Veras RMS, Ramalho GLB, et al. Automatic microaneurysm detection and characterization through digital color fundus images［C］ // Proceedings of the II Workshop on Computational Intelligence. Salvador: eScholarship, 2010: 45-50.
［16］Hatanaka Y, Inoue T, Okumura S, et al. Automated microaneurysm detection method based on doublering filter and feature analysis in retinal fundus images［C］ // 25th International Symposium on ComputerBased Medical Systems.Rome: IEEE, 2012: 1-4.
［17］Narasimhanm K, Neha VC, Vijayarekha K. An efficient automated system for detection of diabetic retinopathy from fundus images using support vector machine and bayesian classifiers［C］ // International Conference on Computing, Electronics and Electrical Technologies. Kumaracoil: IEEE, 2012: 964-969.
［18］高玮玮, 沈建新, 王玉亮. 基于数学形态学的快速糖尿病视网膜病变自动检测算法［J］. 光谱学与光谱分析, 2012, 32(4): 760-764.
［19］Soille P. Morphological Image Analysis: Principles and Applications［M］// Berlin: SpringerVerlag, 1999: 170-171.
［20］Sanchez CI, Hornero R, Lpez MI, et al. A novel automated image processing algorithm for detection of hard exudates based on retinal images analysis［J］. Med Eng Phys, 2007, 30(3): 350-357.
［21］Rosenman J, Roe CA, Cromartie R, et al. Portal film enhancement: Technique and clinical utility［J］. Int J Rdaiat Oncol Biol Physics, 1993, 25(2): 333-338.
［22］British Diabetic Association. Retinal Photography Screening for Diabetic Eye Disease［R］. London: British Diabetic Association, 1997.
［23］Javitt JC, Canner JK, Frank RG, et al. Detecting and treating retinopathy in patients with typeⅠdiabetes mellitus. A healthy policy model ［J］. Ophthalmology, 1990, 97(4): 483-494.
[1]
Song Youyi1Lei Baiying1 He Liang2. 基于超像素和支持向量机的阴道细菌自动检测[J]. 中国生物医学工程学报, 2015, 34(2): 204-211.
