Journal of Biomedical Imaging and Bioengineering

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.
Reach Us +1 (202) 780-3397

Review Article - Journal of Biomedical Imaging and Bioengineering (2021) Volume 0, Issue 0

Automated drusen detection from oct images using genetic algorithm.

Age Related Macular Degeneration (ARMD) is emerging cause of blindness in elderly peoples. ARMD is caused by drusens those are yellow or white spots on the retina. There are basically two imaging technologies widely used by ophthalmologist for retinal analysis, Fundus and Optical Coherence Tomography (OCT). Fundus which is 2D technology and it is lack in detecting early sign of ARMD.OCT Imaging technology is a 3D technique that helps to detect any change in retinal layer structure even in early stage. In this paper we are using OCT images to detect the drusen. Three major steps, image segmentation, feature extraction and classification has been used to diagnose the ARMD from OCT images. Genetic Algorithm (GA) is an optimization algorithm, which produced optimal feature values. In this paper we proposed Genetic Algorithm based optimize feature selection approach for automated diagnosis of ARMD. The overall proposed architecture of this system consists of image segmentation, Histogram Oriented Gradient (HOG) and Local Binary Pattern (LBP) based feature extraction techniques. As Principal Component Analysis (PCA) is used to reduce the dimensionality. This system used the PCA to obtained eign vector values which are further used in Genetic Algorithm. Then we used the Genetic Algorithm for the feature selection, which have produced very optimize value of the features. For classification purpose, we have trained these feature values through multiple classifiers. The proposed system achieved the, overall accuracy 92.5% while sensitivity, specificity are 90.9%, 95.5%. Our proposed Genetic Algorithm produced efficient results with very effective computational time and also with less space as compared to existing approaches. There are two major contributions of this paper. First a detailed survey along with the taxonomy of ARMD is presented and then an optimization algorithm (GA) is used.

Author(s): Samina khalid*

Abstract PDF

Get the App