基于多维特征向量及ANN技术的色彩传递算法  

New color transferring algorithm based on multi-dimensional eigenvector and ANN searching technology

作　　者：滕秀花[1] 陈昭炯[2] 叶东毅[2] 

机构地区：[1]福建工程学院计算机与信息科学系,福建福州350014 [2]福州大学数学与计算机学院,福建福州350002

出　　处：《计算机应用》2006年第12期2866-2868,共3页journal of Computer Applications

基　　金：福建工程学院科研发展基金资助项目(GY-Z0543);福建省自然科学基金(A0410013);福州大学科技发展基金项目(2002-XQ-23)

摘　　要：灰度图像色彩传递可以实现对灰度图像的自动上色处理。分析了目前色彩传递的经典算法——Welsh算法,指出了该算法存在的两大不足之处,提出了一种包含高频纹理信息的向量式像素点描述方式,并结合最近邻域搜索算法(ANN),提出了一种新的色彩传递算法,弥补了传统算法的不足,使灰度图像的彩色化效果有了较大改进。Transferring Color to Grayscale Image is to adjust chromatic channels of grayscale image automatically. The classic algorithms of color transferring named Welsh was analyzed, and its two main disadvantages were pointed out. After introduced a new multi-dimensional eigenvector for pixels with high frequency texture information, a new color transferring algorithm was proposed, which utilizing ANN searching technology. The new algorithm fixes the two main disadvantages of classic algorithm and greatly improves the transferring result.

关 键 词：色彩传递 高频纹理信息 多维特征向量 最近邻域搜索算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]