A detailed description for pruning used in the inference algorithm. 

The proposal A has three children proposal B1, B2 and B3.  The lines between B1, B2, B3 is L1 (B1 and B2), L2 (B1 and B3) and L3 (B2 and B3)

1. Energy function. The threshold is set using RECALL of each node (Recall = 95%). (For And Node and Leaf Node)
2. The number of proposal of each node. Each node has at most 200 proposals after NMS. (For And Node and Leaf Node)
3. The scale and rotation of each proposal.  (For And Node and Leaf Node)
   1. The range of ratio is between 0.5 to 1.5.
   2. The range of rotation is from -PI/3 to PI/3.
   3. Leaf Node: Only using rotation.
4. The difference of scale and rotation between the proposal and its children proposal. (For And Node)
   Scale:   1.5
   Rotation:  PI/4
   1. If its children proposals are from leaf nodes, only Rotation is used.
5. Difference of scale and rotation between these lines and children proposals are used for pruning. For example, the difference between B1 and L1, the difference between B2 and L1 and so on. (For And Node)
   Scale: 1.5
   Rotation: PI/4
   1. If its children proposals are from leaf nodes, only Rotation is used.
   2. The pruning doesn't impact the performane, however, it will speed up the algorithm by 2 seconds.
6. The scale and rotation of  L1, L2, L3 (For And Node) 
   1. The range of ratio is between 0.5 to 1.5.
   2. The range of rotation is from -PI/3 to PI/3.