STTR Phase II: Improving Privacy and Security in Biometrics

这一小型企业技术转让(STTR)第二阶段项目的目标是在生物群方面取得根本性进展--密码保护、隐私增强、指纹和基于面部的技术。该项目将开发原型，以支持商业应用中的Beta测试，并进行大规模的政府测试。开发工作引入了这一概念，并将开发/演示双向生物识别验证，从而传感器和数据库都能收到匹配确认。这对于远程/基于Web的生物识别使用至关重要，并通过卡上匹配解决方案提高了安全性和保密性。它将开发一种新的Biotope，使用但从不存储从现有数据库或潜在指纹无法获得的多光谱数据，提供可持续的、不可欺骗的安全身份令牌，以编码形式匹配并在每笔交易中发生变化。它将探索通过“负”细节点和基于主成分分析的特征增强来提高准确率。为了提高现有基于细节点的算法和硬件的重复利用率，这项工作试图开发一种细节点到细节点的映射方法，并具有与现有生物区相同的安全/隐私保护。对于基于人脸的生物识别，该项目开发了新的多视角方法，用于复杂非结构化环境中非合作主体的基于人脸的验证。此外，该项目解决了隐私保护问题，采用了一种仍支持隐私保护图像存储以进行欺诈起诉的不可搜索技术，并将通过改进远程教育和其他应用的在线验证来扩展持续验证领域的其他研究工作。更广泛的影响始于其独特的重点，即同时提高隐私和安全性，而不是以两者为代价。在公民感到他们的隐私只是为了安全的承诺而被交易的时候，这一努力是对隐私的投资。该项目将把基础研究转变为与商业合作伙伴的测试。它直接解决了其他研究人员所说的导致生物识别领域预测的商业增长与实现的商业增长之间存在永久差距的原因。它将增强生物识别，提供“可撤销”和交易唯一性，以支持基于生物识别的商业，而无需担心网络钓鱼、黑客攻击或内部访问。例如，该项目将影响远程教育市场，重点是提高对有需要的人进行国家培训的有效性，同时保护他们的隐私和尊严。预计的结果还为护照/身份证提供了可能性，使个人能够证明自己的身份，而不允许其他人使用这些数据来搜索他们。它将支持基于智能卡的解决方案，允许进行生物特征验证的匿名交易。它解决了经常被忽视的生物识别困境，即今天生物识别的广泛部署最终可能会增加身份盗窃，并限制明天的生物识别安全价值。

This Small Business Technology Transfer (STTR) Phase II project aims to make fundamental advances in Biotopes -- cryptographically secured privacy-enhanced fingerprint and face-based technologies. The project will develop prototypes to support beta testing in commercial applications and pursue large-scale government testing. The development effort introduces the concept and will develop/demonstrate bi-directional biometric verification, whereby both the sensor and DB receive match confirmation. This is critical for remote/web-based biometric usage and improves security and privacy with match-on-card solutions. It will develop a new Biotope which uses, but never stores, multi-spectral data not obtainable from existing databases or from latent prints, providing a sustainable non-spoofable secured identity tokens that match in encoded form and change on every transaction. It will explore improving accuracy with 'negative' minutiae and PCA-based feature enhancements. To improve reuse of existing minutia-based algorithms and hardware, the effort seeks to develop a minutiae-to-minutiae mapping approach with the same security/privacy protection of existing Biotopes. For face-based biometrics, the project develops new multi-view approaches for face-based verification from non-cooperative subjects in complex unstructured environments. Additionally, the project addresses privacy protection, with a non-searchable technology that still supports a privacy-protecting image-storage for fraud prosecution, and will extend other research work in the area of continuous verification by improving the online verification for distance education and other applications.The broader impact starts with its unique focus on simultaneously improving privacy and security rather than trading one for the other. At a time when citizens feel their privacy is traded for the mere promise of security, this effort is an investment in privacy. The project will transition fundamental research into testing with commercial partners. It directly addresses reasons that other researchers have said cause the perpetual gap between predicted and realized commercial growth in biometrics. It will enhance biometrics, providing 'revocability' and transactional uniqueness to support biometric-based commerce without fear of phishing, hacking or insider access. The project will impact the distance education market for example, by focusing on improving effectiveness of state training for those in need while protecting their privacy and dignity. The projected outcomes also open the potential for passports/IDcard that allow individuals to prove their identity without allowing others to use that data to search for them. It will support smart-card based solutions that allow for biometric-verified yet 'anonymous' transactions. It addresses the often overlooked biometric dilemma, that wide-spread deployments of biometrics today may ultimately increase identify theft and also limit biometrics security value tomorrow.