In this paper, a unique approach of retina recognition system has been proposed. Principal component analysis based dimensionality reduction technique has been used here and discrete hidden Markov model (DHMM) has been applied to classify the retinal images. The retinal images are taken from DRIVE retina database. Then some image pre-processing techniques such as image scaling, median filtering-based noise reduction and edge-detection techniques have been used to ready the retinal image for the system. Features are extracted and reduced using principal component analysis-based dimensionality reduction technique. These reduced features are fed to the DHMM learning algorithm to create the learned retinal template. Finally, DHMM recognition model has been used to recognize the input pattern. Experiments and performance analyses are performed according to the DRIVE retinal dataset.Keywords: Retina recognition, principal component analysis, discrete hidden markov model, feature extraction, dimensionality reductionCite this Article: Islam Md. Rabiul. Retina recognition system using PCA-based discrete hidden markov model. Journal of Network Security. 2015; 3(1): 28–32p.
