Shearlet-Based Feature Extraction for the Detection and Classification of Age-Related Macular Degeneration in Spectral Domain Optical Coherence Tomography Images

Age-related macular degeneration (AMD) is an eye disease that usually affects central vision in people older than 50 years owing to accumulation of fluid in the macular region of the retina. Optical coherence tomography (OCT) is an imaging modality that is being widely used nowadays for the detection of abnormalities in the eye. In this work, a shearlet transform–based method is proposed for automated detection of AMD. The 2-dimensional horizontal slices of spectral domain OCT imaging data are used as input images. Images are first converted to gray scale and denoised using bilateral filter. Denoised images are decomposed by applying shearlet transform and 10 textural features are extracted from the cooccurrence matrices of high-frequency transform coefficients. Based on these features, the OCT images are classified as normal or AMD using support vector machine and k-nearest neighbor classifiers. Results obtained using shearlet-based features are compared with that of wavelet transform–based features. Best results are obtained when shearlet-based features are classified using support vector machine.
Source: Journal of Clinical Engineering - Category: Medical Devices Tags: Feature Articles Source Type: research