改进颜色融合的医学图像彩色化技术

Medical image colorization based on improved color fusion

彩色化后的医学图像能清晰体现患者病灶信息有利于医患沟通。提出改进颜色融合的医学图像彩色化方法,首先利用基于KNN的图像前背景区分算法,强化病灶区域的边界信息;然后以此为约束条件,只需提供简单的着色输入;最后将边界能量引入颜色融合方法,得到较好的着色结果。着色图像保持了原图的灰度信息不变,增加了彩色标记图像的颜色和真实感。实验结果表明,该算法具有较高的精确度,可有效地应用于医学图像彩色化处理。

Coloring grey-level images has the advantage of highlighting the suspected regions,which helps with the communication between doctors and patients. This paper proposed a new medical image colorization approach via an improved color fusion scheme. Firstly,it used KNN-based foreground / background segmentation to strengthen the outline. Secondly,it annotated the image with only a few user specified color scribbles. Finally,it introduced a color fusion algorithm to obtain a better color appearance for the medical image. The resulting image preserved the chromatic information of the source image and retained the original luminance of the colored image. In experiments,the algorithm has a high accuracy and the results demonstrate a good potential for practical applications of the proposed algorithm in the medical image processing.