Question about 'reconstructFloorplan'.
Opened this issue · 3 comments
The results of the model training output already have the rendering of the icon and the room. The information such as the corner has been identified. Why do you need reconstructFloorplan, is the logic and model here still relevant?
模型训练产出的结果就已经有了icon和房间的渲染图，墙角等信息已经被识别出来了，为什么还需要reconstructFloorplan，这里边的逻辑和模型还有关系吗？
Sorry for the confusion. ReconstructFloorplan is the IP module which optimize the connection between wall corners. It serves as a post-processing of the network outputs.
Sorry for the confusion. ReconstructFloorplan is the IP module which optimize the connection between wall corners. It serves as a post-processing of the network outputs.
What do you mean by IP module?
I meant the Integer Programming post-processing mentioned in the paper.