Haochen

It would be really nice if this could work with Nuclide.

@apchenstu could you elaborate a bit more on why it cannot handle background? Thank you!!

This pull request is not merged.

https://stackoverflow.com/questions/6022302/how-to-apply-unmerged-upstream-pull-requests-from-other-forks-into-my-fork Maybe this can help.

I created a separate branch just for this pull request. It's called rpn-testing

This codebase does not support Fast-RCNN training (the one you described) or RPN-only training. Both require restructuring model_builder.py. I did not do this pull request since it is better for...

Some slight modifications will make it possible. You might need to sprinkle a few more changes here and there down the road. But the spirit is true.

Hi, this would be very useful if the loop is compute-intensive to run, and one only wants to test it for a few before jumping out to test the subsequent...

The high precision of depth value cannot be contained in PNG format. That's why we used npy array. What are your particular use cases? Visualization?

PNG sacrifices precision; hence it's not stored that way.