摘要
为了从时延红外人脸图像中提取出更加稳定的血流特征,提出1种新的基于生物热传模型的人脸血流图建模方法,并应用于人脸识别。基于生物热传模型中经典的Pennes生物传热方程通过离散化建模,得到对应的血流图。然后,基于红外人脸图像低分辨率的特点,对得到的血流图采用经典主成分分析PCA(principal component analysis,PCA)+Fisher线性鉴别分析(Fisher linear discriminant analysis,FLD)的方法进行人脸识别。实验结果表明:本研究提出的血流计算方法充分考虑到血流之间的关联性,可以得到更加稳定的血流特征,用于人脸识别,在不同环境温度的情况下(对于时延数据)具有较强的鲁棒性。
To get stable biological features from the time-delay infrared face, a new construction method of blood perfu- sion was proposed based on bio-heat transfer, which could applied to face recognition. According to the classic bio-heat transfer equation （Pennes equation）, the blood perfusion rates in different positions were computed based on the discrete method. Then, due to the low-resolution of the infrared images, feature extraction method ＇ PCA （ principal component analysis） ＋ FLD （ Fisher linear discriminant analysis） ＇ was chosen to get the principle features in the blood perfusion image. The experiment results showed that the blood perfusion rate proposed was stable, its application in infrared face recognition was robust to the impact of the environment temperature.
出处
《山东大学学报（工学版）》
CAS
北大核心
2013年第5期1-5,12,共6页
Journal of Shandong University（Engineering Science）
基金
国家自然科学基金资助项目(61201456)
江西省自然科学基金资助项目(20132BAB201052)
