Iris recognition using neural network based on VHDL prototyping
Mamun Bin Ibne Reaz, M.S. Sulaiman, Faisal Mohd-Yasin, Tan Ai Leng · 2004
This paper describes the development of a synthesizable VHDL model of a neural network application to recognize irises for user verification. Iris has unique features to be used as a biometric signature due to its speed, simplicity, accuracy, and applicability. Fourier transform is used to transform the image into data that can feed into the network. A two layer neural network is chosen with the input layer having three neurons, hidden layer having two neurons and the output layer having one neuron. The sigmoid function has been used as the activation function. The Iris recognition offers better alternative way of user verification. The Iris preprocessing had done in Matlab software and tested.