How to use the scenecut with only the RGB-D images?

[ywcmaike]

Hi, gays! I'm interested in your research scenecut. Now I want to extract the plane in the image, I found your model fit me. I have read the papers and code. I was confused that how to use the RGB-D images to scenecut. I found your code only give the demo that need the image, the hierarchies and the pointcloud.
Can you tell me how to generate the hierarchies and the pointcloud? (or give me the example to generate the hierarchies and pointcloud in matlab or c++ or python).
I read the papers that you use the network cob(Convolutional Oriented Boundaries) to get the boundaries, and then use the hierarchies to segment tree, and the optimise tree cut results in the object segmentation. I want to implement the code(may be python or C++) just as the papers said. and I need your help.
Can you give me the whole code and the whole process when you have conduct the research?

thanks!
@trungtpham

[ShrutheeshIR]

Hey @ywcmaike . I'm intending to do something similar to what you have described. Were you successful? If so, please let me know. Thanks!