Uncovering Vein Pattern using Generative Adversarial Network

Vein distribution is important in medical treatments. It could also be used for identity authentication(1). As a basic part of our body, the blood vessel has the merits of universality and distinctiveness. However, vein patterns are usually not visible in color images, which carries significant limitation. To address this limitation, we proposed a deep-learning-based method. Our method can uncover vein distributions from color images, help relieving pains to patients and widening the application scenarios of vein patterns. Experimental results showed that the proposed method has reliable performance and robustness in varying environments.