Retina Images Based Human Eye Retinal Biometric System

A human eye retinal based biometric system is a pattern recognition system that recognizes a human's eye retinal image on the basis of features derived from a specific physical characteristic that the person possesses.There are so many biometric technologies;one of them is biometric pattern retina recognition which is used for identification of the peoples.Biometric patter retina recognition that is a programmed procedure which provides exact recognition of a single person by gaining the inner body image the retina of an individual who is willing must cooperate in such a way in which duplication is very difficult.The human retina is a very tinny tissue which comprises of slight cells.It is compulsory to use images of human retina for the improvement of security.In order to avoid crime and counterfeiting,the human eye retina has a unique style and function,so it is very valuable and protected to the greatest extent.In this research work of mine,I used convolutional neural network(CNN)feature extraction and long-term short-term memory(LSTM)classification model or neural network(NN)to project a biometric recognition system based on human eye retina images.For this reason,retinal images data is augmented by using several pre-processing techniques;the data is augmented because the dataset contain only one image per person.Smoothness,noise addition,rotation,scale,mirror,sharpness,color shifting and contrast management these basic approaches are used to augment the retina images data.The proposed work was simulated by using MATLAB(Matrix Laboratory)simulation tools.In simulation based results,the four models on the basis of Epoch,LSTM Units and batch-size have been evaluated with each other.Each model has a different accuracy expressed as a percentage,and compared with the existing three models,model four shows significant accuracy.