Peter Anderson

Check under the training heading of the Readme https://github.com/peteanderson80/bottom-up-attention#training

No we didn't use the class labels or the bbox. I did some initial experiments like that but performance didn't change much. Mostly we used the boxes just for visualization.

This is working as intended. The image is a skybox / cubemap not an equirectangular panorama, so it's actually 6 different images representing the faces of a cube and they...

Assuming the make command runs without error, it looks like python is not looking in the correct place for the MatterSim package. This looks like a python issue. The python...

Are you running on a machine with an x-server? If it's a server than the answer is probably no. If so, try one of the off-screen rendering options: https://github.com/peteanderson80/Matterport3DSimulator#rendering-options-gpu-cpu-off-screen

Hi @jotaf98, I enjoyed your MapNet talk, very cool. Depth / segm is a priority so we will get this done asap. I think we will want to do some...

We have added depth outputs. See the README for details, these are based on the undistorted depth images which are not necessarily perfectly aligned to the skyboxes. I haven't got...

My expectation is that projecting cnn features to the ground plane would be fine. The location of cnn features in the image is quite coarse anyway (with large receptive fields),...

Yes it would be possible. There is no demo because I haven't ever tried it. I'd be really interested to see how it works. You would need to modify the...

Yes, it is based on LMDB