Then propose two fast methods for training and execution, Selective Gain Computation algorithm and Sparse FeatureTree, accordingly.
然后分别从训练和执行两个方面提出快速方法:选择性增益计算方法和稀疏特征树方法。
3
The vessel skeleton was recursively tracked to generate the featuretree and the tree was traversed consequently to pick up the branch points and end points.