这是一篇来自于 微软和卡耐基的文章，iccv2015best paper 以及其对应的 材料supplemental materials
主要贡献
- A novel approach that unifies classification trees with the representation learning functionality known from deep convolution networks, by training them in an end-to-end manner.
- Provide a globally optimal strategy to estimate predictions taken in the leaves.
决策树示意图
决策树示意图
