Reproduce Co-Fusion paper result on car4 and room4 about co-fusion HOT 2 OPEN

Hi,

I also tried to run it on these two sequences recently. I used the example command line arguments from dataset tools as well. The ATE of car4 and room4 I obtained are 2.7cm and 1.4cm respectively. Although they are still not as good as the ones reported in the paper, they look more promising than yours.

Here's the arguments I used:

./CoFusion -basedir path/to/car4-full/ -dir colour -depthdir depth_noise  -run -ep -colorprefix Color -depthprefix Depth -cal calibration.txt -segMinNew 0.008   -crfSmooth 1   -crfPos 0.51 -exportdir estimate -maskprefix nomask -es -keep

Could you share more details about the command you used to run it? For example, did you specify the calibration parameters using -cal. This would be a possible reason you obtained such a large ATE.

from co-fusion.