欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
为了改善运动估计中传统光流法处理的图像存在平滑性不好，边界信息丢失等缺点，提出了一种新的图像运动估计方法。该方法加入图像结构不变假设，降低光照变化的影响，并利用K—Mean聚类法对图像分割矫正，使得运动估计结果在保持平滑的基础上，更好地保留图像的边界信息。实验结果表明，该算法比传统光流法在运动边界问题的处理上更有好的改善。
英文摘要：
A novel approach is proposed for improving the loss of the smoothness and boundary information of the optical flow estimation for images in traditional motion estimation algorithms. Due to these problems, the assumption of the fixed image structure is added and the influence of illumination changes is reduced. Furthermore, the K-Mean clustering algorithm is used to restore more boundary information from the input images without losing extra smoothness in the stored images. The experiments show that the proposed method has a better result on the motion boundary estimation issues.
同期刊论文项目
同项目期刊论文
Alteration in scaling behavior of short-term heartbeat time series for professional shooting athlete
A new measure to characterize multifractality of sleep electro- encephalogram, Chinese Science Bulle
期刊信息
位置：