The complex structure of an iris image result in the difficulty of iris representation especially for iris images of insufficient quality.it has been found that several accurate iris recognition algorithms employ multiscale techniques, which provide a well suited representation for iris data. existing methods either assume that iris images are of sufficient quality or devote much effort to accurately segment them. To address this problem especially when images are noisy or distorted this book proposes a new multiscale approach as a pre-processing step to efficiently localise the iris followed by a new normalisation scheme based on both an optimum unwrapping reference to capture the data and dynamic size to avoid inherent interpolation inefficiencies which cause loss of information. A feature extraction technique, which is based on a combination of some multiscale feature extraction techniques, is also proposed and evaluated.In addition, optimisation techniques are proposed to reduce the effect of eyelids and eyelashes.The experimental results have shown that the proposed system could be used for personal verification in an efficient and effective manner
Il n'y a pas encore de discussion sur ce livre
Soyez le premier à en lancer une !
Les meilleurs albums, romans, documentaires, BD à offrir aux petits et aux plus grands
Il n'est pas trop tard pour les découvrir... ou les offrir !
Inspirée d’une histoire vraie, cette BD apporte des conseils et des solutions pour sortir de l'isolement
L’écrivain franco-vénézuélien Miguel Bonnefoy poursuit l’exploration fantasmagorique de sa mémoire familiale...
