用画笔工具在训练图像中拾取那些可能是肤色的像素，，并分别记录下这些像素在RGB和HSV颜色空间的颜色值，根据大量的样本值，确定肤色在两个颜色空间三通道的阈值，用所确定的阈值寻找测试图像中的肤色区域。一种简单的可视化这一过程的方法是将所有非肤色像素指定为黑色，肤色像素指定为白色。(Use the brush tool to pick up those may be the color pixels in the training images, and records the pixels in the HSV color space and RGB color values, according to a large number of samples, determine the color in the two color spaces of three channel threshold, looking for the test in the color image region with the determined threshold. A simple way to visualize this process is to specify all non - skin pixels as black, and the skin color pixels are white.)