The Palmprint region segmentation is an important step in Palmprint recognition problem. A comparative study of two segmentation algorithm namely Square-based ROI detection method and Circle-based ROI detection Method are studied and implemented. Firstly, the Square-based ROI detection method uses the Binarization of Palmprint and extract the Boundary, after this detecting key points and establishing a coordination system with extracting the central part as a fixed size square. The circle based ROI detection method uses a 2D rectangular coordinate system with the y-axis formed by the line connecting k1 and k2 and the x-axis is the perpendicular line originating from the midpoint of k1 and k2. And the largest inscribed circle is extracted by searching the centre of line joining k1 - k2 and line perpendicular to it, using four directions. Palmprint region segmentation is realized based on these two methods. The experiment results show that the algorithm is accurate and effective.
Published In : IJCSN Journal Volume 5, Issue 2
Date of Publication : April 2016
Pages : --
Figures :15
Tables : --
Publication Link : Comparative Study of ROI Extraction of Palmprint
Milind E.Rane : SSBT COE&T, North Maharashtra University Jalgaon, India
Umesh S Bhadade : SSBT COE&T, North Maharashtra University Jalgaon, India
Palmprint ROI detection, square based method, circle based method
The Palmprint segmentation algorithms are studied, the segmentation based on Square-based ROI detection method and Circle-based ROI detection Method are implemented. The simulation results show that Palmprint segmentation results are better, and gives the area of valid Palmprint i.e. Region of Interest which have important features for Palmprint recognition. These two methods the overcome the shortcomings caused by palm rotation and translation. The circle based method gives the larger area under ROI as compared to Square-based method which gives more information in circle based method. But the radius of circle based method is not fixed so different area is captured during ROI detection where as in square based method the fixed size area is extracted. Also the interference caused due to the boundary region is more in circle based method compared to square based method.
[1] Zhong Qu, Zheng-yong Wang, “Research on Preprocessing of Palmprint Image Based on Adaptive Threshold and Euclidian Distance”, Sixth International Conference on Natural Computation (ICNC 2010)IEEE, 2010
[2] H. Kipsang Choge, T. Oyama, Stephen, “Circle-based Region-of-interest segmentation method for Palmprint recognition”, ICROS-SICE international joint conference, 2009.
[3] The PolyU Palmprint Database (version 2.0); http://www.comp.polyu.edu.hk/~biometrics
