Radon and Multiwavelet-Based Compact Feature Vector Generation for Gender Identification from Iris

Person identification from iris biometrics has gained good popularity over the other biometrics because of its high accuracy. It would be more advantageous if iris biometrics provides person attributes such as gender, ethnicity, and age besides identity. This paper explores a new method to extract gender from the iris image. In this paper, a new feature extraction technique based on the combination of 2D Multiwavelet and Radon transform is proposed. Statistical features are computed to form a final feature template. Support vector machine is trained to classify gender from the computed feature vector. Experimental results obtained, show that the proposed method outperforms all earlier methods and it can be implemented for gender classification from iris image.