基于皮肤毛囊光学相干层析（OCT）成像的三维断层生物特征识别新方法

With the rapid development of Internet and information technology, information security becomes an important issue. There is an increasing demand for human identification and feature inference, especially based on the individual's unique physiological or behavioral attributes, so called biometric technology, which gains widespread attention and becomes a hot research topic. However, as most of biometric technology is limited to a one-dimensional or two-dimensional recognition, the recognition rate is vulnerable to the external environment, forgery and other factors. The three-dimensional cross-sectional biometric technology is a more accurate and reliable techniques with more information and less susceptible to counterfeiting. In this proposal, we focus on a new biometric feature---the hair follicles, via a novel three-dimensional optical coherence tomography (OCT) imaging system, which enables three-dimensional, cross-sectional, high-resolution, wide field of view, large depth imaging. The introduction of OCT technology breaks through the bottleneck that most of current three-dimensional biometric are only based on the analysis of three-dimensional contour imaging. The OCT system is also expected to realize automatic extraction and measurement of follicle features, such as density, shape, distribution and spectral characteristic. A human hair follicle feature database will be established and analysed to perform the feature matching tests and group feature studies. This proposal aims to develop a novel three-dimensional cross-sectional biometric technology for a new biometric feature---human hair follicles with a novel three-dimensional OCT system, and improve the effectiveness of biometric technology eventually.

随着互联网和信息技术的快速发展，人们对信息安全、个体识别和个体特征推断的要求日益提高，基于个人独特生理、行为特征的生物特征识别技术受到广泛重视和研究。然而，目前大多数生物特征识别技术还局限于一维或二维识别，识别率容易受环境、伪造等因素影响，三维生物特征识别技术鲁棒性更好，信息更丰富，伪造难度更大，是一种更准确可靠的技术。本项目以人体皮肤毛囊这一新型生物特征为研究对象，通过研制一套新型三维光学相干层析(OCT)成像系统，突破目前大多三维生物特征识别技术只能进行三维轮廓成像的瓶颈，实现对人体皮肤毛囊的三维、高分辨、宽视场、大成像深度的断层成像和光谱成像，并自动提取和测量毛囊分布形态、光谱等特征参数，建立小型体皮肤毛囊图像特征数据库，进行特征匹配测试和群体特征规律研究。本项目旨在通过新型三维OCT成像技术，初步建立一种针对人体皮肤毛囊的三维断层生物特征识别新方法，提高生物特征识别的效果和水平。

生物特征识别技术近年来受到广泛重视和研究。皮肤毛囊是一种新型的生物识别特征，但是目前的皮肤毛囊识别技术还仅仅局限于在皮肤外表面进行毛囊二维成像分析和识别，尚无法获取毛囊深度方向信息以实现三维断层分析。本项目采用光学相干层析（OCT）技术对人体皮肤毛囊进行三维高分辨断层成像，通过提取毛囊的深层多维特征，初步实现基于皮肤毛囊的三维生物特征识别分析。本项目研制了一套新型三维频域OCT系统，研发了一套手持式样品扫描装置，实现了平台式和手持式多模式扫描成像，同时进一步提升了系统各项性能指标，达到高分辨、超宽视场、大成像深度的总体要求，获取了多种不同法庭科学物证样品和皮肤毛囊的高质量三维OCT图像。研发了毛囊OCT图像自动识别、分割、测量算法，针对人脸7个不同部位，提取了11个新型量化特征，采集了20个志愿者共计140份中国人皮肤毛囊活体样本数据，建立了一个小型皮肤毛囊OCT图像数据库，提取并测量了相应特征参数，对这些特征参数进行了统计学差异分析，研究了毛囊特征的部位差异、个体差异，初步探索了毛囊特征与性别、年龄等属性特征之间的关联。本项目通过有效利用皮肤毛囊的三维结构信息，有望为案事件中个体识别和特征推断提供新方法新手段，拓展和完善现有的技术方法体系。