人体切片数据集颜色的自动校正方法

目的 校正拍摄数字人切片过程中存在的种种干扰，使各层照片颜色参数趋于一致。方法 提出了一种基于CMYK色卡，对照片在RGB空间进行自动颜色校正的算法。该算法根据色卡中色块分布特点自动提取出每张照片中的色卡灰度部分，对此细分得到15个灰度值，并由此计算出和真实颜色之间的变换函数，根据该函数确定出颜色校正的查找表，在RGB颜色空间通过查表可对亮度和色彩进行校正。结果 利用Matlab 6．5编程实现了该算法，在微机上完成了数字人女婴一号原始数据的颜色校正．每张照片的处理时间为13．475s。结论 应用该算法能对人体切片数据集进行自动快速地颜色校正，校正后数据集层间颜色的一致性得到提高。

An Automatic Color Correction Algorithm for Digital Human Body Sections

Objective To find a new approach to improve the uniformity of color parameters for images data of the serial sections of the human body. Method An auto-color correction algorithm in the RGB color space based on a standard CMYK color chart was proposed. The gray part of the color chart was auto-segmented from every original image, and fifteen gray values were attained. The transformation function between the measured gray value and the standard gray value of the color chart and the lookup table were obtained. In RGB color space, the colors of images were corrected according to the lookup table. Result The color of original Chinese Digital Human Girl No.1(CDH-G1) database was corrected by using the algorithm with Matlab 6.5, and it took 13.475 s to deal with one picture on a personal computer. Conclusion Using the algorithm, the color of the original database is corrected automatically and quickly. The uniformity of color parameters for corrected dataset is improved.