| 基于虚拟曝光融合的光照自适应人脸图像增强方法 |
| |
| 作者姓名： | 靳晓缘 徐望明 伍世虔 |
| |
| 作者单位： | 武汉科技大学信息科学与工程学院,湖北 武汉,430081;武汉科技大学机器人与智能系统研究院,湖北 武汉,430081,武汉科技大学信息科学与工程学院,湖北 武汉,430081;武汉科技大学冶金自动化与检测技术教育部 工程研究中心,湖北 武汉,430081;武汉科技大学机器人与智能系统研究院,湖北 武汉,430081,武汉科技大学信息科学与工程学院,湖北 武汉,430081;武汉科技大学机器人与智能系统研究院,湖北 武汉,430081 |
| |
| 基金项目： | 国家自然科学基金资助项目(61775172)；湖北省教育厅科研计划项目(D20191104). |
| |
| 摘 要： | 为了提高复杂光照条件下的人脸识别性能,提出一种基于虚拟曝光融合的光照自适应人脸图像增强方法。该方法针对检测到的人脸图像,计算其平均亮度,并与经由统计学习得到的正常光照人脸图像的亮度区间进行比较,判定其光照水平,分为低光照、高光照或正常光照,然后基于相机响应模型采用虚拟曝光融合的方法对低光照和高光照人脸图像进行光照增强处理,此过程循环迭代处理直到增强的人脸图像平均亮度达到正常水平。这种光照自适应增强后的人脸图像可作为输入无缝接入现有的人脸识别算法中,从而改善人脸识别系统性能。在Extended Yale B和CMU_PIE人脸图像数据库上的实验结果表明,该方法能有效提高复杂光照下的人脸识别率。
|
| 关 键 词： | 人脸识别 图像增强 光照自适应 曝光融合 相机响应模型 光照水平 |
| 收稿时间： | 2019/9/24 0:00:00 |
An illumination-adaptive face image enhancement method using virtual exposure fusion |
| |
| Authors: | Jin Xiaoyuan Xu Wangming and Wu Shiqian |
| |
| Institution: | College of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan 430081, China;Institute of Robotics and Intelligent Systems, Wuhan University of Science and Technology, Wuhan 430081, China,College of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan 430081, China;Engineering Research Center for Metallurgical Automation and Detecting Technology of Ministry of Education, Wuhan University of Science and Technology, Wuhan 430081, China;Institute of Robotics and Intelligent Systems, Wuhan University of Science and Technology, Wuhan 430081, China and College of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan 430081, China;Institute of Robotics and Intelligent Systems, Wuhan University of Science and Technology, Wuhan 430081, China |
| |
| Abstract: | To improve the face recognition performance under complex illumination conditions, an illumination-adaptive face image enhancement method using virtual exposure fusion is proposed. Firstly, the average brightness of a detected face image is calculated and its illumination level is classified as low, high or normal by comparing it with the brightness interval obtained via statistical learning from the face images with normal illumination.Then a virtual exposure fusion method based on camera response model is used to enhance the face image which has low light or high light. This procedure is iterated until the average brightness of the processed image reaches the normal level. The final enhanced face image can be seamlessly integrated to the existing face recognition algorithm as an input to improve the system performance. Experimental results on Extended Yale B and CMU_PIE face image databases indicate that the proposed method can effectively improve the accuracy of face recognition under complex illumination. |
| |
| Keywords: | face recognition image enhancement illumination self-adapation exposure fusion camera response model illumination level |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《》浏览原始摘要信息 |
| 点击此处可从《》下载免费的PDF全文 |
|
