掌静脉识别预处理部分包括ROI提取和图像增强。ROI 提取一般包括四个步骤:手掌图像二值化、手掌轮廓线提取、利用相关算法定位关键点并建立坐标系和提取合适的 ROI。
1 ROI提取之基于关键点定位
【硕士论文】基于人手姿态估计的掌纹感兴趣区域提取算法研究(动态视频和静态图像)
论文1:双目形态学扩展法。Outside Box and Contactless Palm Vein Recognition Based on a Wavelet Denoising
论文2: Z2VDA(Zig-Zag valley detection)算法。Preliminary Work on Rotation-Invariant Algorithms for Contactless Palm Vein Biometrics
论文3: CHVD(competitive hand valley detection) 算法
论文4:8邻域链码:论文:一种快速的手掌轮廓特征点提取方法
论文5:径向分布函数法(radial distribution function,RDF)


2 ROI提取之基于内切圆
论文1:Hand Geometry Analysis by Continuous Skeletons
3 ROI提取之基于深度学习
论文1:Finger Vein Recognition Based on Multi-Task Learning.
算法:利用多任务学习方法共同优化ROI提取任务和特征提取任务。
论文2:Palmprint Recognition in Uncontrolled and Uncooperative Environment.
论文3:【2023】YOLOV7-TMV3: Lightweight non-contact palm print region of interest segmentation model
论文4:【2024】Complete Region of Interest for Unconstrained Palmprint Recognition
论文5:人脸定位:HyperFace: A Deep Multi-task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition