345ishaan / denselidarnet

DenseLidarNet

Project Title

Generating Dense Lidar Data using cues from monocular image and sparse lidar data.

Kitti Data Visualization

Sparse to Dense Halluncination

Please find the descriptions of the files in the code folder as follows:-

1. ./code/train.py => Main file used to push the trigger for training. Supports flag for train/eval and cuda enable/disable.
2. ./code/vfe_layer.py => VFE Layer and DenseLidarNet Model in Pytorch
3. ./code/voxelize.py => Creates a voxel map for sparse lidar map and generates masks and indices.
4. ./code/dataloader.py => Pytorch DataLoader customized for DenseLidarNet.
5. ./code/chamfer_loss.py => Chamfer Loss Computation in Pytorch