| 李伟红,陈龙,龚卫国.单样本快速人脸不变特征提取方法[J].光电子激光,2014,(3):558~564 |
| 单样本快速人脸不变特征提取方法 |
| Fast facial descriptor extraction for single image based face recognition |
| 投稿时间：2013-06-24 |
| DOI： |
| 中文关键词: 单样本 人脸识别 积分图 |
| 英文关键词:single image face recognition integral image |
| 基金项目:国家自然科学青年基金(61105093)和重庆市重点科技攻关(CSTC2012-YYJSB40001)资助 项目 (重庆大学 光电工程学院,光电技术及系统教育部重点实验室,重庆 400044) |
|
| 摘要点击次数: 1891 |
| 全文下载次数: 200 |
| 中文摘要: |
| 本文提出一种快速人脸特征描述(FFD,fast facial descriptor)算法和基于权重的人脸图像相似度分数匹配策略, 以解决加速鲁棒性特 征(SURF,speed up robust features)在描述单样本人脸特征时出现的特征点分布不均 匀和光照变化鲁棒性差的问 题。首先通过重构积分图来增加位于细长边缘的特征点的数量；为了减少冗余特征,提 出两幅训练图像对应特征点间的 区别度概念,对训练样本中的特征点进行稀疏化；然后,根据人脸各区域对识别结果贡献度 的不同对人脸各区域赋予不同权 重,并根据加权计算人脸图像的相似度分数得出识别结果。在AR、Yale B和CMU PIE标 准人脸数据库及真实身份证人 脸库上进行了单样本人脸识别实验。结果表明,本文算法对具有光照、遮挡和表情变化的单 样本人脸识别有很好的鲁棒性, 耗时仅为0.042s；与目前典型的特征描述算法相比,本文算法的识 别率最高可提升65%；虽 然真实身份证人脸库中人脸图像受实际环境因素影响较大,但本文方法也可提高30%的识别率。 |
| 英文摘要: |
| In this paper,a fast facial descriptor (FFD) algorithm and a face ima ges similarity score matching strategy are proposed to improve the uneven distribution of features and poor illumination robus tness o f speed up robust features (SURF) in describing features for single image face recognition.The integral image is reconstructed to increase the number of features on the lathy edge of face image to alleviate the problems of SURF.In order to reduce the number of r edundant features,the conception of distinction between the corresponding features is introduced to extract sparse features on t raining images.According to the contribution to recognition rate,face image regions are weighted to calculate the similarity sc ore between face images to conduct the final result. Experiments of single image based face recognition are conducted on AR,Yale B,CM U PIE databases and real-life ID card image database.The results indicate that the proposed method is robust to single imag e based face recognition under illumination,occlusion and expression changes,and the running time is only 0.042s.Compared wi th some other recent feature description methods, the proposed method can increase the recognition rate by up to 65%.Although the images in real-life ID card image database are greatly affected by actual environmental factors,the recognition rate is increa sed by 30%. |
| 查看全文 下载PDF阅读器 |
| 关闭 |
|
|
|
|
|