A Novel Approach for Human Identification Based on Vein Structure in Sclera

Abstract

The vein structure within the sclerotic coat, the white and opaque outer protecting covering of the attention, is
anecdotally stable over time and distinctive to every person. As a result, it's well suited to use as a biometric for human
identification. A few researchers have performed sclerotic coat vein pattern recognition and have according promising,
however low accuracy, initial results. sclerotic coat recognition poses many challenges: the vein structure moves and
deforms with the movement of the eye and its close tissues; pictures of sclerotic coat patterns square measure typically
defocused and/or saturated; and, most significantly, the vein structure within the sclerotic coat is multi-layered and has
complicated non-linear deformation. The previous approaches in sclerotic coat recognition have treated the sclerotic coat
patterns as a one-layered vein structure, and, as a result, their sclerotic coat recognition accuracy is not high. In this
thesis, we have a tendency to propose a new methodology for sclerotic coat recognition with the following contributions:
1st, we have a tendency to develop a color-based sclerotic coat region estimation theme for sclerotic coat segmentation.
Second, we have a tendency to design a Dennis Gabor rippling based mostly sclerotic coat pattern improvement
methodology, associate degreed an adaptive thresholding methodology to emphasize and binarize the sclerotic coat vein
patterns. Third, we have a tendency to projected a line descriptor based mostly feature extraction, registration, and
matching methodology that's scale-, orientation-, and deformation-invariant, and will mitigate the multi-layered
deformation effects and tolerate segmentation error. It's through empirical observation verified mistreatment the UBIRIS
and IUPUI multi-wavelength databases that the projected methodology will perform correct sclerotic coat recognition. In
addition, the recognition results square measure compared to iris recognition algorithms, with terribly comparable results.
Keywords- Sclera vein recognition, Feature extraction, sclera feature matching, sclera matching