segmentation-driven-pose How to make a mesh label for semantic segmentation during training

How to make a mesh label for semantic segmentation during training

Open 571502680 opened this issue 5 years ago • 5 comments

Jul 19 '19 07:07 571502680

2D reprojections of the mesh vertex according to the pose can tell you ground truth segmentations.

Jul 19 '19 09:07 YinlinHu

But the paper uses the Semantic Separation Grid of S*S，if i use 2D reprojections of the mesh vertex, and its size is the shape of pictures!

Jul 20 '19 03:07 571502680

Just use the label of the nearest pixel, to generate a rough segmentation.

Jul 20 '19 20:07 YinlinHu

Could you share the train.py?

Jul 24 '19 07:07 571502680

We will release an enhanced version of our method with the training code. (maybe in months)

Jul 30 '19 09:07 YinlinHu