Medical Biometrics Enabling Security Technologies (MBEST)

越来越多的安全解决方案基于生物识别技术，用于身份认证、安全监控系统、数据通信和访问控制。然而，常用的生物识别模式，如指纹、虹膜和人脸，都很容易被规避、重播和混淆攻击。另一方面，医学生物特征识别包括一组新的模式，如心电(CG)、脑电(EEG)、耳声发射(OAE)、心音(PCG)等，由于它们位于人体内部，因此难以窃取或伪造，因此对这些攻击具有很强的抵抗力。此外，它们提供固有的活性检测和持续身份验证的可能性，因为它们可以每隔几秒提供新的生物识别读数。 这项提议的短期目标是调查医学生物识别管道，包括各种医学生物识别模式的获取、预处理、登记、模板生成、匹配和分类。我们将解决的重要挑战是由于潜在信号的时间依赖性而产生的模板可变性和稳定性。长期目标是建立医学生物识别，无论是作为独立的还是多生物识别组件，作为医疗保健、现场监测和访问控制应用中的实时安全应用的可行解决方案。这将需要利用廉价、实时的DSP平台来实现更高的识别率、更短的采集和处理时间以及更低的成本。最后，将引入生物特征加密技术，以提高隐私。 这项研究为生物识别系统的研究和设计提供了新的思路，并将使人们更好地理解医学生物识别模式及其在可靠的个人认证和识别方面的应用。此外，它还解决了对生物识别系统的开发及其在安全和控制的职业和福利监测部署中的应用具有实际意义的及时问题。

A growing number of security solutions is based on Biometrics recognition technologies for identity authentication, secure monitoring systems, data communication and access control . However, the commonly used biometric modalities such as the fingerprint, iris and face, are venerable to circumvention, replay and obfuscation attacks. On the other hand, medical biometrics which include a novel set of modalities such as the electrocardiogram (ECG), the electroencephalogram (EEG), Oto-acoustic emissions (OAE), the Phonocardiogram (PCG), among others, are robust to these attacks because they are internal to the body and therefore difficult to steal or forge. Furthermore, they offer inherent liveness detection and possibility of continuous authentication since they can provide a fresh biometric reading every couple of seconds. The short term objectives of this proposal is the investigation of the medical biometrics pipeline including the acquisition, pre-processing, enrollment, template generation, matching and classification of various medical biometric modalities. Important challenges that we will address are template variability and stabilization arising from the time dependent nature of the underlying signals. The long term objective is to establish medical biometrics, either as a stand alone or a multi-biometrics component, as viable solution to real time security applications in health care, field monitoring and access control applications. This will require to achieve higher recognition rates, shorter acquisition and processing times and lower cost by utilizing cheap, real time DSP platforms. Finally, biometric encryption technologies will be introduced for increased privacy. The proposed research presents novel ideas in biometric systems research and design and will provide better understanding of medical biometric modalities and their utilization for reliable person authentication and recognition. Also, it addresses timely issues of practical importance to he development of biometric systems and their applications in occupational and welfare monitoring deployments for security and control.