项目名称: 基于距离度量学习和类依赖特征分析的人脸特征提取方法研究
项目编号: No.61201359
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 电子学与信息系统
项目作者: 严严
作者单位: 厦门大学
项目金额: 25万元
中文摘要: 近年来,复杂环境下的人脸识别受到了广泛的研究和关注。如何提取有效的人脸特征是人脸识别的关键问题之一。本项目的研究目标是针对复杂环境下的人脸识别,在距离度量学习和类依赖特征分析级联结构框架的基础上研究有效的鉴别特征提取方法。从大规模数据半正定度量学习、增量学习、相关滤波器和相关滤波器组的设计等方面分别进行深入研究,建立起一套有效的人脸鉴别特征提取方法,从而在一定程度上克服由光照、姿态、表情和年龄等各种内外因素所造成的人脸识别困难。本项目组具有良好的人脸识别研究工作基础。对本项目的研究将进一步推动人脸识别技术的发展。
中文关键词: 人脸识别;鉴别特征提取;距离度量学习;类依赖特征分析;
英文摘要: In recent years, face recognition in an unconstrained environment has received extensive research and attention. How to extract the effective feature is one of the key issues for face recognition. The aim of this project is to study effective discriminant feature extraction methods for accurately recognizing faces in an unconstrained environment. Based on the cascade structure of distance metric learning and class-dependence feature analysis, we focus on the large-scale semidefinite metric learning, incremental learning,the design of correlation filter and correlation filter bank. The project tries to develop effective discriminant feature extraction algorithms for face recognition, so that the recognition difficulties caused by various factors such as illumination, pose, expression, age, can be relatively alleviated. The project team has a good foundation for the face recognition research. This project will bring the further improvement to the face recognition technology.
英文关键词: face recognition;discriminant feature extraction;distance metric learning;class-dependence feature analysis;