Towards the generation of synthetic images of palm vein patterns: A review

With the recent success of computer vision and deep learning, remarkable progress has been achieved on automatic personal recognition using vein biometrics. However, collecting large-scale real-world training data for palm turned out to be challenging, mainly due noise irregular variations included at time acquisition. Meanwhile, existing datasets are usually collected under near-infrared light, lacking detailed annotations attributes (e.g., pose), so influences different have poorly investigated. Therefore, this paper examines suitability synthetic images generated compensate urgent lack publicly available datasets. Firstly, we present an overview research recognition, from basic background knowledge anatomical structure, acquisition, public database, quality assessment procedures. Then, focus state-of-the-art methods that allowed generation vascular structures biometric purposes modeling biological networks with their respective application domains. In addition, review style transfer nature-based algorithms. Afterward, formalize a general flowchart creation database comparing real samples obtain some understanding into development realistic imaging system. Ultimately, conclude by discussing challenges, insights, future perspectives in generating further works.