kielnino

Results 22 comments of kielnino

Performance does not scale well with very large slide shows

I have the same problem.

Added different upsampling-modes

I realised, that I have to update the train script too. I need to get used to the argparse with choices first, so I will update the file as soon...

Added different upsampling-modes

Okay, I think it should work now.

Added different upsampling-modes

@Gouvernathor You are right. I've updated `train.py` and `predict.py`

Added different upsampling-modes

@milesial Is there anything else I can/must adjust to get the PR approved?

Add images with submitted annotations, but empty bounding boxes to the coco dataset

@lfeldb This is also a solution for [heartexlabs/label-studio#1470](https://github.com/heartexlabs/label-studio/issues/1470). But when I use your PR and upload an image in LS and afterwards try to create a COCO-export I get the...

Speed up reconstruction using external initial pose estimates

I would also like to have the ability to insert prior pose informartion. Unfortunately `reconstruct_from_prior` is only suitable to add new images to an existing reconstruction.

Speed up reconstruction using external initial pose estimates

@fabianschenk Sure, but then the PRs must also be accepted at some point. My other PR is waiting for a month now (https://github.com/mapillary/OpenSfM/pull/957) Best, Johannes

Speed up reconstruction using external initial pose estimates

@fabianschenk thank you. I have noticed the slow down but as the project is very mature, it is a pleasure to work with anyway. Thanks for making it public available....

returning norm_coords instead of px_coords in pixelToNormalizedCoordi…

Hey @fabianschenk, can you take a look at this? That looks like an important bug fix to me.