International Journal of Computer Science Issues

More than a traditional journal...

This paper presents a robust multimodal multifeature biometric authentication scheme integrating iris and lip images based on feature fusion. This paper is one of its kinds, as there is no significant work presented till now that involves multimodal and multifeature. Also, the integration of lip and iris for multimodal is unique. The ROI (region of interest) extraction from the input iris image is obtained using Hough circles. As there are no databases available for lip, the viola face detection algorithm is adopted to obtain the required ROI from the face image. Feature extraction involves the process of extracting multiple features such as texture and line. Hough transform is used for the procurement of the line feature extraction. Canny edge detection algorithm is implemented to obtain Hough transform. Texture feature extraction process is carried out using Haralick method. To surmount the restriction of the possible missing modalities, the multiple parallel support vector machines (SVMs) classification strategy is applied. Fusion of two modalities is carried out at the feature level. This work is to study investigation of better alternative verification techniques suitable for fusion of two modalities, as well as fusion of iris and lip feature at an earlier stage.