项目名称: 基于结构特征的人脸识别研究
项目编号: No.61502304
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 计算机科学学科
项目作者: 高盛华
作者单位: 上海科技大学
项目金额: 21万元
中文摘要: 人脸识别具有重要的学术价值和应用价值。本课题将着重研究基于结构特征的人脸识别,尤其是利用深度结构和稀疏结构对人脸进行表示和特征提取。本课题将着重研究以下三个关键技术问题:基于结构特征的人脸关键点对准技术;基于深度结构和稀疏结构相结合的人脸识别技术;基于深度学习的针对注册人脸和待识别人脸分辨率不同的人脸识别技术。通过这三个问题的研究,可以加深对深度学习等相关技术的理解,促进人脸识别技术的发展,推进其在实际中的应用。
中文关键词: 人脸识别;关键点对齐;卷积神经网络;自编码器;稀疏编码
英文摘要: Face recognition is an extremely important task for both research community and real applications. In this project, we will work on face recognition based on the structured features, which are extracted by leveraging the deep characteristics and the sparse characteristics in face representation. Specifically, we will focus on the following key issues, i.e. i) Face alignment based on the structured features of faces; ii) Extracting deep and sparse features for face recognition; iii) Deep learning for face recognition when the resolutions of the enrolled faces and probe faces are different. It is desirable that the study of these tasks helps us to deepen the understanding of deep learning, contribute to the development of the face recognition, and push the applications of face recognition in real problems.
英文关键词: Face Recognition;Face Alignment;Convolutional Neural Networks;Auto-Encoder;Sparse Coding