Extraction of Dynamic Region of Interest (ROI) for Palmprint

Palmprint authentication is a means of personal authentication that uses unique palmprint features. Palmprint is obtained by just scanning the user's palm on the platform of the scanner when scanning is performed. So, it is subjected to various physical disturbances such as variable size, noise and orientation, which will degrade the verification process. Therefore image preprocessing plays an important role in palmprint recognition. Preprocessing is a technique of aligning different palmprint images, obtaining coordinate systems and to segment the central region for feature extraction. The proposed algorithm focuses on extraction of Dynamic Region Of Interest (ROI) from the palmprint image. Most of the existing work uses static region from palmprint not utilizing a significant portion of the palm. Intuitively, the larger area captures more distinctive features when compared to fixed size ROI. The experimentations are performed on the PolyU database to validate the proposed algorithm.