Engineering Commons^™

An Iris Authentication System Based On Artificial Neural Networks, Brenden Velu

Electrical Engineering

An iris authentication system verifies the authenticity of a person based on their iris features. The iris features are extracted through wavelet transform of the isolated iris from modified iris images. A level 5 wavelet decomposition is performed on the images, and the resulting low-frequency wavelet coefficients represent the inputs to the artificial neural network. The artificial neural network reads these features as inputs, and classifies each set of inputs according to their target identity. This authentication system currently classifies up to 10 people. The irises used for classification represent ideal situations with minimum eyelash and eyelid interference.