The result reveals more human-vision-alike segmentation and much less computational cost.

• 实验中得到了比已有算法更接近人眼视觉的分割结果并大大减少了算法的运算量。