摘要
为了提高对眼动跟踪图像校正的精确性和快速性,提出一种基于改进型的图像校正方法。首先将采集到的瞳孔跟踪数据映射到一张图像上,并用重心提取法对图像做去噪处理,然后把整张图像划分成九个均匀的矩形区域,最后在各个矩形区域内,分别用多项式坐标变换方法对注视坐标点进行坐标变换。实验分析表明,该方法能够有效地校正失真图像,运行时间相对于传统算法减少了近2/3,具有良好的实用价值。
To improve the accuracy and speed of the correction of eye tracking image, a kind ot improvement image correction algorithm is discussed. First, the collected eye tracking data is maptcl to an inrago, and the image is removed noise by the methed of extracting center of gravity. Then the whole image is divided into nine uniform rectangular area. Finally in each rectangulararea, polynomial eootrtinate transform mcthed is used to tranfform the gaze coordinates eoordinate. The experimental analysis shows that the algorithm can efl＇ectivaly correct the distortion image, whose maning time compared with the traditional algorithm is reduced by nearly 2/3, and has a good practical value.
出处
《激光杂志》
CAS
CSCD
北大核心
2013年第2期32-34,共3页
Laser Journal
基金
重庆市科技攻关计划基金资助重点项目CSTC2011GBB0028
重庆飓鹰科技有限公司
关键词
眼动跟踪
注视点、图像失真、校正、坐标变换
eye traching
fixation point
image distortion
correction,coordinate transformation
