A study on palm vein image generation using near infrared images and color images based on deep learning

Reo Tokuke; Makoto Takamatsu; Eizaburo Iwata; Makoto Hasegawa

doi:10.1117/12.2541828

Translator Disclaimer

16 October 2019 A study on palm vein image generation using near infrared images and color images based on deep learning

Reo Tokuke, Makoto Takamatsu, Eizaburo Iwata, Makoto Hasegawa

Author Affiliations +

Proceedings Volume 11205, Seventh International Conference on Optical and Photonic Engineering (icOPEN 2019); 112050Q (2019) https://doi.org/10.1117/12.2541828
Event: Seventh International Conference on Optical and Photonic Engineering (icOPEN 2019), 2019, Phuket, Thailand

Abstract

Palm veins images are taken by a near infrared camera and a common digital camera. It is possible to generate clear vein images using near-infrared rays, but this is not easy using a color image. A near infrared palm vein image for training data and a color image were arranged in a dataset pair; the datasets were learned using a deep learning method called pix2pix. Clear palm vein images close to the near infrared image were generated using our proposed method. Our experimental results showed that we can take a palm vein image using only a general digital camera without a near infrared camera. Realization of palm veins authentication on an inexpensive smartphone is expected.

Citation Download Citation

Reo Tokuke, Makoto Takamatsu, Eizaburo Iwata, and Makoto Hasegawa "A study on palm vein image generation using near infrared images and color images based on deep learning", Proc. SPIE 11205, Seventh International Conference on Optical and Photonic Engineering (icOPEN 2019), 112050Q (16 October 2019); https://doi.org/10.1117/12.2541828

ACCESS THE FULL ARTICLE