用libsvm来实现图像分割，测试图片用的亦是25cases和40cases中的那个littleduck测试图片。主体程序思想为25cases中的代码过程，改进之处为可以让用户利用ginput来提取背景的样本点和前景（待分割出来的目标）的样本点作为训练样本，而不需实现指定背景和前景的样本点，也不用额外的小软件来查看某点的RGB值，ginput即可。-With libsvm to image segmentation, test picture is also used in 25cases and 40cases that littleduck test picture. The main idea for the program code is in the process of 25cases, improvements to allow users to take advantage of ginput to extract the sample point background and foreground (to be carved out of the goal) of the sample points as training samples, without the need to achieve the specified background and foreground sample point, do not have the little extra software to view the RGB values of a point, ginput can be.