Study of Heterogeneous Dorsal Hand Vein Recognition Based on Multi-device

摘要

The effectiveness of dorsal hand vein recognition technology depends on image quality. The problem of dorsal hand vein image heterogeneity is becoming increasingly prominent in the era of big data. In the multi-device acquisition process, image ROI size, contrast, sharpness, position shift and image rotation are the main parameters of image heterogeneity. In order to explore the effects of different parameters on the image multi-device recognition, we adjusted 5 quality parameters of dorsal hand vein image individually first of all. Then, we used different recognition algorithms for experiment, and quantitatively analyzed the effect of different parameters on the heterogeneous dorsal hand vein image recognition by the improvement of recognition rate. Finally, the method of multi-parameter adjustment was proposed to improve recognition rate.