论文部分内容阅读
提出一种将新颖的眼睑弹性模板与霍夫变换相结合实现眼睛特征的自动提取的方法，即自动定位眼球并自动提取眼睛轮廓．与其它用弹性模板提取眼睛特征的方法相比，本方法可以处理对比度较低、下眼睑模糊的图像，提取特征准确度更高，且改进的能量函数和保守能量下降算法大大简化了经验参数的选择．经过百余幅眼睛图像的实验处理，验证了该方法的有效性．
A novel method of automatically extracting eye features by combining the novel eyelid elasticity template with Huff transform is proposed to automatically locate eye contours. Compared with other methods of extracting eye features by elastic template, this method can process images with low contrast and lower eyelid, and extract features with higher accuracy. The improved energy function and conservative energy reduction algorithm greatly simplify the empirical parameters select. After more than 100 eyes of the experimental image processing to verify the effectiveness of the method.