摘要
为提高面部疲劳状态的识别效果,提出了一种融合全局特征和局部特征的面部疲劳特征表示方法.该方法将离散余弦变换(discrete cosine transform,DCT)和独立元分析(independent component analysis,ICA)技术以及Gabor变换相结合,通过融合全局独立DCT特征和局部动态Gabor特征得到最终的面部疲劳特征表示.基于前人自建的疲劳图像序列库进行了实验,结果表明该方法提取的疲劳特征更加具有鉴别力.
To improve the facial fatigue state recognition effect, a facial fatigue features representation method of fusing the global and local features was proposed. This method combined the discrete cosine transform （DCT）, independent component analysis （ICA） technology and Gabor transformation, and obtained the final facial fatigue features representation through fusing the global independent DCT features and local dynamic Gabor features. The experiments based on the previous self-built fatigue image sequences show that the fatigue features extracted by this method are more discriminative.
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2013年第1期63-69,共7页
Journal of Beijing University of Technology
基金
国家自然科学基金资助项目(60973057)
关键词
疲劳分析
特征融合
离散余弦变换
独立元分析
GABOR小波
fatigue analysis
features fusion
discrete cosine transform （DCT）
independent component analysis （ICA）
Gabor wavelet
