Charles R. Qi

Hi @fferroni Yes, we are working on the KITTI rectified camera coordinate system. If you want to test on another dataset, you need to verify that the axis directions are...

Hi @fferroni No problem -- glad it works. we'd better oversample the actual points by repeating the existing points (by randomly choices) -- putting them to zero introduces an un-wanted...

HI @cjhyy7190 You have to set the axis direction right otherwise the prediction will be meaningless. In default the camera coordinate system has Y pointing downward, so it's very different...

hi @kl456123 can you be more specific on the indent problem?

Hi @VenkatesanSoundirarajan What you need are corresponding image patches (in 2D boxes) and frustum point clouds. In our case with the KITTI data, we need camera intrinsics to know which...

Hi @kentaroy47 that's very interesting. Can you share a bit more on your dataset/use case?

Thanks @kentaroy47 for the extra info. I think if the background is subtracted, then the voting seems shouldn't be too hard. It might be helpful to visualize the ground truth...

@chinacui It's possible. If the scene contains mostly background points, votes may not have very high density, so the difference between boxnet and votenet could be small. A variation of...

Hi @amiltonwong are you running with ssh session? there might be some issue with remote access. if you are running locally, it might take a few seconds to display the...

Hi @zemahran sorry for the late response. But you can refer to [this function](https://github.com/charlesq34/frustum-pointnets/blob/aaa660554cf18e449392eb743bf6afbdca4a8d44/kitti/kitti_object.py#L124), basically you just need to get the 3D coordinates of the 3D box and project them...