Gorilla-Lab-SCUT/frustum-convnet

I have quseiton about frustum-convnet

Gironote opened this issue · 2 comments

Gironote commented

I wonder what deep learning model rgb_detection.txt file was created using.
image

Screenshot from 2020-09-11 16-16-14.png
Or, Did you save the information in Kitty datasets in a different form again?

zhixinwang commented

The rgb_detction.txt comes from F-PointNets. You can refer to the details on the supplements of F-PointNets. The format of this file is image_path(or image ID), category ID, confidence, x1, y1, x2, y2.

Gironote commented

@zhixinwang
As far as I know, f-convnet requires 2D object detection deep learning.
Does this program bring the 2d detection results from the kitti label?