项目名称: 一种非均匀光照和局部遮挡环境下维吾尔族人脸识别算法研究
项目编号: No.61462082
项目类型: 地区科学基金项目
立项/批准年度: 2015
项目学科: 其他
项目作者: 伊力哈木·亚尔买买提
作者单位: 新疆大学
项目金额: 46万元
中文摘要: 针对人脸识别存在鲁棒性和实时性问题,尤其在新疆特殊的地理环境下,存在光照强和非均匀性的特点,以及维吾尔族人具有脸部遮挡的风俗习惯,这就会带来维吾尔族人脸识别的鲁棒性和实时性瓶颈等难题,亟待解决。本课题以新疆维吾尔族人脸在非均匀光照和局部遮挡环境下的识别为研究目标,以算法的鲁棒性和实时性为评价体系,提出一种非均匀光照和局部遮挡环境下维吾尔族人脸识别算法研究。重点解决维吾尔族人脸面部最具辨识性特征的提取、非均匀光照变化的情况下维吾尔族人脸的识别、维吾尔族人脸面部遮挡情况下的特征点的提取等一系列科学问题。通过本课题研究,实现具有良好实时性和鲁棒性的非均匀光照和局部遮挡情况下维吾尔族人脸识别算法,以此开展维吾尔族人脸识别方面的应用研究。其成果在身份识别、公安照片搜索系统、社会稳定等领域有着良好的应用前景,同时也能够促进对新疆地区的人脸识别方面的研究,为构筑平安新疆有着重要的学术意义和应用价值。
中文关键词: 非均匀;局部遮挡;维吾尔族;人脸识别
英文摘要: Face recognition in Xinjiang is often difficult due to special geographic environments, such as light intensity and non-uniformity, and customs of the Uighurs who are often with faces obscured. These have be the bottleneck problems in achieving robustness and real-time processing in Uighurs face recognition, which needs to be specially addressed. This project focuses on Uighur face recognition under non-uniform illumination and partial eclipsing, aims at robustness of algorithms and real-time processing and accordingly proposes a algorithms. The algorithm focus on solving discriminative feature extraction under the conditions such as light intensity and non-uniformity illumination and masked faces. The obtained methods are expected to be applied in identification and photo search system. The research can be a great promotion of face recognition research in the Xinjiang region, peace construction, and of great academic significance.
英文关键词: non-uniform;Partial shade;The uighurs;Face recognition