Building a verification system based on the finger vein image is a hard problem since it goes through many steps including vein detection, features extraction and classification. Also, dealing with finger misalignment makes it even harder. In this work three methods for vein detection is applied: Maximum Curvature, Repeated Line Tracking and Wide Line Detector. Two approached for features extraction is used: Histogram of Oriented Gradient (HOG) and Orientation Map. ELM and SVM are used as classifier. - View it on GitHub