Error when testing the trained model for eggbox on LMO

[xuliwalker]

Hi, when I evaluate your trained model for eggbox on LMO, I encountered the following error. How can I fix this issue? Thanks!

[suyz526]

Hi,

I'm not aware of this. Maybe you can try with a different OpenCV version? I used OpenCV 4.5.2

[xuliwalker]

Thanks for your reply! I changed the opencv-python version to 4.5.2.52, but it still did not work. When I changed the parameter minimum_point_number of the function pyprogressivex.find6DPoses from 6 to 4, it worked. I wonder whether this change would affect the model evaluation performance. Thanks!

Hi,

I'm not aware of this. Maybe you can try with a different OpenCV version? I used OpenCV 4.5.2

[suyz526]

Sorry for the unclear answer. I meant the version of opencv-python, which can be installed with pip install opencv-python==4.5.2.54 .

[xuliwalker]

Thanks for your reply! I changed the opencv-python version to 4.5.2.54, but it still did not work. When I changed the parameter minimum_point_number of the function pyprogressivex.find6DPoses from 6 to 4, it worked. I wonder whether this change would affect the model evaluation performance. Thanks!

[suyz526]

I think it will not affect the final results.

If the matching is too less, the estimated pose has very less chance of being a correct pose. In other words, even if there is no error report, the estimated pose is bad.