A machine learning approach to medical image classification: Detecting age-related macular degeneration in fundus images

Age-Related Macular Degeneration (AMD) is a dangerous, chronic, and progressive illness that mostly affects people over 60 years old. This disease is related to the appearance of drusen: deposits of extracellular material located in the macular region. One way to effectively and non-invasively pre-diagnose AMD is by detecting the presence of drusen in fundus images. In this work we propose a new method that combines Digital Image Processing, Mathematical Morphology and a robust and powerful Machine Learning model: a Support Vector Machine (SVM). The enclosed macular region is subjected to a contrast enhancement method, followed by the application of basic morphological operations. We use invariant moments as the features of the processed image. The resulting vector is classified by an SVM as positive or negative for drusen. The proposed method is able to discriminate between healthy and afflicted cases with a classification accuracy that outperforms many well-regarded state-of-the-art methods.

[1]  U. Rajendra Acharya,et al.  Application of different imaging modalities for diagnosis of Diabetic Macular Edema: A review , 2015, Comput. Biol. Medicine.

[2]  György Kovács,et al.  A self-calibrating approach for the segmentation of retinal vessels by template matching and contour reconstruction , 2016, Medical Image Anal..

[3]  S. Edward Rajan,et al.  Retinal vessel segmentation employing ANN technique by Gabor and moment invariants-based features , 2014, Appl. Soft Comput..

[4]  Deepti Mittal,et al.  Automated detection and segmentation of drusen in retinal fundus images , 2015, Comput. Electr. Eng..

[5]  Elijah Blessing Rajsingh,et al.  An empirical study on optic disc segmentation using an active contour model , 2015, Biomed. Signal Process. Control..

[6]  Mark A. Hall,et al.  Correlation-based Feature Selection for Machine Learning , 2003 .

[7]  A.D. Hoover,et al.  Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response , 2000, IEEE Transactions on Medical Imaging.

[8]  Aaron Y. Lee,et al.  Deep learning is effective for the classification of OCT images of normal versus Age-related Macular Degeneration , 2016, bioRxiv.

[9]  Roberto Rosas-Romero,et al.  A method to assist in the diagnosis of early diabetic retinopathy: Image processing applied to detection of microaneurysms in fundus images , 2015, Comput. Medical Imaging Graph..

[10]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[11]  Qiang Chen,et al.  Automated drusen segmentation and quantification in SD-OCT images , 2013, Medical Image Anal..

[12]  Oscar Camacho Nieto,et al.  Clasificador de Heaviside , 2015 .

[13]  Ming-Kuei Hu,et al.  Visual pattern recognition by moment invariants , 1962, IRE Trans. Inf. Theory.

[14]  Mohamed Amine Guerroudji,et al.  A new approach for the detection of mammary calcifications by using the white Top-Hat transform and thresholding of Otsu , 2016 .

[15]  Bernhard E. Boser,et al.  A training algorithm for optimal margin classifiers , 1992, COLT '92.

[16]  Gwénolé Quellec,et al.  Exudate detection in color retinal images for mass screening of diabetic retinopathy , 2014, Medical Image Anal..

[17]  U. Rajendra Acharya,et al.  Diagnosis of retinal health in digital fundus images using continuous wavelet transform (CWT) and entropies , 2017, Comput. Biol. Medicine.

[18]  Philippe Burlina,et al.  Comparing humans and deep learning performance for grading AMD: A study in using universal deep features and transfer learning for automated AMD analysis , 2017, Comput. Biol. Medicine.

[19]  Ian H. Witten,et al.  The WEKA data mining software: an update , 2009, SKDD.

[20]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .