Over the past few decades, hand-based multimodal biometrics systems have achieved significant attention because of their high security, accuracy, and anti-counterfeiting. Various hand physiological biometric modalities have been explored for identity authentication, i.e., fingerprint, finger knuckle print, palmprint, palm vein, and dorsal hand vein traits. This study provides a comprehensive review focusing on the interface of different hand biometric traits and presents an overview of hand-based multimodal biometrics methods. The framework of this paper is divided into three main categories. Firstly, we introduce the characteristics of four levels of hand-based biometrics in detail. Following this, several typical image capturing devices and image preprocessing techniques of various hand-based biometrics are reviewed. Moreover, existing publicly available and widely used hand-based multimodal biometrics databases are then summarized. Subsequently, the hand-based multimodal biometrics methods are categorized into sensor-level fusion, feature-level fusion, score-level fusion, rank-level fusion, and decision-level fusion. Additionally, the recent hybrid fusion-based and deep learning-based hand multimodal biometrics approaches are analyzed and discussed. Furthermore, we conduct a performance analysis of the abovementioned algorithms from the recent literature. At last, challenges, trends, and some recommendations related to hand-based multimodal biometrics are drawn to give some research directions.

中文翻译：

---

基于手部的多模态生物识别融合：综述

在过去的几十年里，基于手的多模态生物识别系统因其高安全性、准确性和防伪性而受到广泛关注。已经探索了用于身份认证的各种手部生理生物识别模式，即指纹、指节纹、掌纹、手掌静脉和手背静脉特征。本研究对不同手部生物识别特征的界面进行了全面综述，并概述了基于手部的多模态生物识别方法。本文的框架分为三个主要类别。首先，我们详细介绍了四个级别的手部生物识别技术的特点。接下来，回顾了几种典型的图像捕获设备和各种基于手的生物识别技术的图像预处理技术。此外，还总结了现有的公开可用且广泛使用的基于手的多模态生物识别数据库。随后，基于手的多模态生物识别方法被分类为传感器级融合、特征级融合、分数级融合、等级级融合和决策级融合。此外，还分析和讨论了最近基于混合融合和基于深度学习的手部多模态生物识别方法。此外，我们根据最近的文献对上述算法进行了性能分析。最后，提出了与基于手的多模态生物识别技术相关的挑战、趋势和一些建议，以给出一些研究方向。