【摘要】 为了提高颜色传递的速度与精确性,改善构成不相似图像之间的色彩传递效果,对传统的Welsh颜色传输算法进行了改进.首先颜色空间由RGB空间转移到以视觉感知为基础的lαβ空间下,再按照一定标准将源图像局部的平均值和标准差组成特征矢量群.根据图像局部区域具有相似特征的特点,对源图像与目标图像局部矢量特征进行抽样匹配计算.实验结果表明,该算法具有计算精确、传递速度快,受图像构成相似与否干扰小的优点.更多还原

【Abstract】 A modified color transfer method based on Welsh’s algorithm is introduced to raise the color transfer speed and accuracy,and improve the effects between dissimilar images.Firstly,RGB signals are converted to visual perception-based color space lαβ,and feature vector is formed by the mean values and standard deviation according to a certain standard.Local vector of swatches between the target and the source image are matched and calculated according to the principle that the corresponding local areas have similar characteristics.The experimental results demonstrate that the proposed algorithm is of high computation precision,fast transfer speed and strong disturbance rejection between dissimilar images.更多还原