| 基于多尺度稀疏表示的面部疲劳识别 |
| |
| 引用本文： | 牛耕田,王昌明,孟红波.基于多尺度稀疏表示的面部疲劳识别[J].计算机科学,2016,43(8):282-285, 291. |
| |
| 作者姓名： | 牛耕田 王昌明 孟红波 |
| |
| 作者单位： | 南京理工大学机械工程学院 南京210094,南京理工大学机械工程学院 南京210094,南京理工大学机械工程学院 南京210094 |
| |
| 基金项目： | 本文受国家高科技研究发展计划(863计划)(2012AA061101),高维信息智能感知与系统教育部重点实验室(南京理工大学)开放基金(3092013012205),高等学校博士学科点专项科研基金(20133219110027)资助 |
| |
| 摘 要： | 针对疲劳驾驶严重威胁道路交通安全的问题,提出了一种基于多尺度稀疏表示的面部疲劳识别算法。该算法首先通过Gabor小波获取面部多尺度多方向的疲劳特征；然后采用2D-PCA方法对提取的特征进行降维处理,提高算法的执行效率；最后通过稀疏表示的方法构造疲劳的超完备字典并完成疲劳识别。实验在自建的疲劳数据库中完成,结果显示所提算法的疲劳识别率达到94.5%,具有一定的可行性。
|
| 关 键 词： | 疲劳识别 Gabor小波 2D-PCA 超完备字典 稀疏表示 |
| 收稿时间： | 2015/9/22 0:00:00 |
| 修稿时间： | 2/2/2016 12:00:00 AM |
Fatigue Recognition Based on Spare Representation |
| |
| NIU Geng-tian,WANG Chang-ming and MENG Hong-bo.Fatigue Recognition Based on Spare Representation[J].Computer Science,2016,43(8):282-285, 291. |
| |
| Authors: | NIU Geng-tian WANG Chang-ming and MENG Hong-bo |
| |
| Affiliation: | School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China,School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China and School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China |
| |
| Abstract: | In order to solve the traffic safety problems caused by fatigue driving,a method based on sparse representation was proposed to detect the fatigue through face image.In this method,first,Gabor wavelets are used to extract multi-scale and multi-orientation features.At the same time,considering the execution efficiency of the algorithm,2D-PCA is used to reduce the dimension of features.Finally,based on sparse representation theory,the over-complete dictionary of fatigue is constructed and fatigue is identified.The proposed method was tested on the self-built database.Experimental results show the effectiveness of the proposed method,and the fatigue recognition rate reaches 94.5%. |
| |
| Keywords: | Fatigue recognition Gabor wavelets 2D-PCA Over-complete dictionary Sparse representation |
|
|
点击此处可从《计算机科学》下载全文 |
|
