为了减少alpha-expansion算法的计算量，本文在标号为alpha的像素向其它像素膨胀的过程中，先隔离非alpha类间的联系，而只考虑alpha类与非alpha类之间的关系，从而避免alpha-expansion算法需要构造辅助结点的问题，减少了s/t图中边的数目，提高了算法的计算效率。因放松了非alpha类间的关系对alpha膨胀的约束，使得算法可以更容易得跳出能量函数的局部极小点而获得更优的分割结果。-In order to reduce alpha- expansion algor

The ultimate goal of machine vision is image understanding - the ability not only to recover image structure but also to know what it represents. By de¯ nition, this involves the use of models which describe and label the expected structure o

Four crucial issues are considered by the proposed HoAL: 1) unlike binary cases, the selection granularity for multilabel active learning need to be fined from example to examplelabel pair 2) different labels are seldom independent, and label correla

Four crucial issues are considered by the proposed HoAL: 1) unlike binary cases, the selection granularity for multilabel active learning need to be fined from example to examplelabel pair 2) different labels are seldom independent, and label correla