search

epiception / CalibNet

[DEPRECATED] Self-Supervised Extrinsic Calibration using 3D Spatial Transformer Networks
https://epiception.github.io/CalibNet/
MIT License
215 stars 57 forks source link

velo_to_cam Matrix is necessary ? #5

Closed AI-liu closed 6 years ago

AI-liu commented 6 years ago

dataset_build_color.py velo_to_cam = np.vstack((np.hstack((velo_to_cam_R, velo_to_cam_T)), np.array([[0,0,0,1]])))

velo_to_cam Matrix is calibration result? Is it necessary ?

Thanks!

epiception commented 6 years ago

For the ground truth generation, yes. But its never explicitly provided to the network during training to supervise upon.

oaix commented 5 years ago

For the ground truth generation, yes. But its never explicitly provided to the network during training to supervise upon. @epiception If we get the ground truth of calibration result, it's unnecessary to train this network. The ground truth is what we need.