Compression standards in finger vein recognition

Abstract

The impact of applying three ISO/IEC still image compression standards to finger vein recognition accuracy is assessed. While JPEG is not competitive for low bitrates, JPEG 2000 and JPEG XR turn out to perform well for different types of template generation techniques. PSNR as a generic measure for image quality is found not to be suited as a predictor for finger vein recognition performance under compression artifacts.