Daniel Barath
Daniel Barath
I merged the python wrapper into the main repo (https://github.com/danini/magsac.git). Hopefully, it solved this problem as well.
Hi. That should be the estimated model parameters.
Dear @xwasco , Thanks :) I can, but I will most likely not have time to do that before ECCV deadline. After that, I certainly will. Best, Daniel
Yeah, I didn't add multi-line fitting to the repository. I will upload it later this week.
I added multi-line fitting here: https://github.com/danini/progressive-x/blob/master/examples/example_multi_lines.ipynb I haven't performed any parameter tuning on line fitting problems, so the results probably are suboptimal.
Hi, Thanks! There is an example for line fitting at https://github.com/danini/progressive-x/blob/master/examples/example_multi_lines.ipynb You can use any data as input there (e.g., star5, star11, or you custom one).
That is I think a windows file. You can probably remove from the include list.
Ah yeah, I see that Church is missing from 2DGS too. The F1 scores I get with the provided code are: - Barn 0.4813 - Meetingroom 0.0601 - Ignatius 0.6712...
Okay, after checking carefully, only the alignment of Truck seems off. I guess I used the wrong trajectory file there, have to double-check. The others look correctly aligned, however, a...
Hey, I got back to this now. So I ran without distortion loss and also tried without the appearance embedding, but neither seems to solve the problem. On Meetingroom, the...