A Multi-view Fusion Method for Finger Vein Recognition
Weishi Li, Sihang Xu, Miao Zhang, Yong Jun Peng · 2023
With the wide application of biometrics and the development of multi-mode fusion, how to obtain more accurate recognition technology by combining multiple characteristic information needs to be solved. Based on the finger vein recognition technique, this paper proposes a multi-view method of high-dimensional feature layer fusion at the feature level by canonical correlation analysis, which can improve the accuracy of recognition results. The fusion method of finger veins and fingerprint was demonstrated in the experiment of two-stream network. The effectiveness of the multi-view fusion method has also been demonstrated on several common baseline data sets. The results presented are competitive with the most advanced results in the literature.