LUTBIO multimodal biometric database

The LUTBIO database provides a comprehensive resource for research in multimodal biometric authentication, featuring the following key aspects:

- Extensive Biometric Modalities: The database contains data from nine biometric modalities: voice, face, fingerprint, contact-based palmprint, electrocardiogram (ECG), opisthenar (back of hand), ear, contactless palmprint, and periocular region.

- Diverse Demographics: Data were collected from 306 individuals, with a balanced gender distribution of 164 males and 142 females, spanning an age range of 8 to 90 years. This diverse age representation enables analyses across a wide demographic spectrum.

- Representative Population Sampling: Volunteers were recruited from naturally occurring communities, ensuring a large-scale, statistically representative population. The collected data encompass variations observed in real-world environments.

- Support for Multimodal and Cross-Modality Research: LUTBIO provides both contact-based and contactless palmprint data, as well as fingerprint data (from optical images and scans), promoting advancements in multimodal biometric authentication. This resource is designed to guide the development of future multimodal databases.

- Flexible, Decouplable Data: The biometric data in the LUTBIO database are designed to be highly decouplable, enabling independent processing of each modality without loss of information. This flexibility supports both single-modality and multimodal analysis, empowering researchers to optimize, combine, and customize biometric features for specific applications.

✅ Data Availability: To facilitate early access, we are initially releasing sample data from 6 individuals. Upon publication of the paper, the full dataset will be made available.

🥸 Important Notice: Please read the data collection protocol of the LUTBIO dataset carefully before use, as it is essential for understanding and correctly interpreting the dataset. Thank you.

😎 Good news! Our paper has received revision comments, and we are carefully making revisions based on the feedback. We appreciate the reviewers and the editor for their efforts.🥰🥰🥰

🤠 We conducted further research based on the LUTBIO database and proposed AuthFormer: An Adaptive Multimodal Biometric Authentication Framework for Secure and Flexible Identity Verification. The code is available at https://github.com/RykerYang/Authformer.git. Our work is currently under review for an international conference, where we are dealing with a malicious reviewer. We are in the process of responding and appealing—wish us luck! 🍀🍀🍀

LUTBIO数据库为多模态生物特征认证领域的研究提供了全面的资源支持，其核心特点如下：

- 丰富的生物模态类型：该数据库涵盖9种生物特征模态的数据，分别为语音、人脸、指纹、接触式掌纹、心电图（electrocardiogram, ECG）、手背、耳朵、非接触式掌纹以及眼周区域。

- 多样化的人口统计属性：数据采集自306名个体，性别分布均衡，其中男性164名、女性142名，年龄跨度为8至90岁。这种广泛的年龄覆盖范围支持开展多人口统计维度的分析研究。

- 具有统计代表性的人群采样：志愿者均来自自然社区，保障了样本规模与统计代表性，所采集的数据涵盖真实场景下的各类变异情况。

- 支持多模态与跨模态研究：LUTBIO同时提供接触式与非接触式掌纹数据，以及指纹数据（含光学成像与扫描两种采集方式），可推动多模态生物特征认证领域的技术进步。本数据库旨在为后续多模态数据库的开发提供参考范式。

- 灵活可解耦的数据结构：LUTBIO数据库中的生物特征数据采用高度解耦的设计，可在不损失信息的前提下对各模态数据进行独立处理。这种灵活性既支持单模态分析，也支持多模态分析，使研究人员能够针对特定应用场景优化、融合与定制生物特征。

✅ 数据可用性说明：为支持早期研究探索，本团队先期公开6名个体的样本数据。待论文正式发表后，将完整发布全部数据集。

🥸 重要声明：使用本数据集前，请务必仔细阅读LUTBIO数据集的数据采集协议，这对于正确理解与解读数据集至关重要。感谢您的配合。

😎 好消息！我们的论文已收到审稿修改意见，目前正基于反馈意见进行细致修订。衷心感谢审稿人与编辑的辛勤工作🥰🥰🥰

🤠 基于LUTBIO数据库，我们开展了进一步研究并提出了AuthFormer：一种面向安全灵活身份验证的自适应多模态生物特征认证框架。其代码已开源至https://github.com/RykerYang/Authformer.git。目前本工作正处于国际会议审稿阶段，期间我们正应对一名恶意审稿人的意见，正在进行回复与申诉——祝我们好运！🍀🍀🍀

搜集汇总

背景与挑战

背景概述

LUTBIO multimodal biometric database是一个多模态生物特征认证研究资源，包含来自306名不同年龄和性别个体的九种生物特征数据，如声音、面部、指纹等。该数据库支持单模态和多模态分析，数据设计灵活可分离，并遵循严格的数据收集和隐私保护协议。