Closed CK-Sung closed 4 years ago
Hey @CK-Sung, you could have a look at the infer function, which returns the semantic estimates with probabilities over all classes with the shape of (_n_classes, _img_h, _img_w).
Thank you so much !
@CK-Sung Thank you for following our work, and I will appreciate it a lot if you could star or fork our repo too.
Hi, I am running the rangenet++ and also the OverlapNet.
The OverlapNet can be run with depth, normal, and intensity. But, I would like to add the label data as well.
So, I tested the RangeNet++ to generate the label data with pre-trained weight, but the code only shows the visualization results.
How can I save the label data which is the result of the RangeNet++ to use in the OverlapNet?