godspeed1989 / ACMNet

Adaptive Context-Aware Multi-Modal Network for Depth Completion

ACMNet

This the Pytorch implementation of our work on depth completion.

S. Zhao, M. Gong, H. Fu and D. Tao. Adaptive Context-Aware Multi-Modal Network for Depth Completion. PAPER

Environment

1. Python 3.6
2. PyTorch 1.2.0
3. CUDA 10.0
4. Ubuntu 16.04
5. Opencv-python
6. pip install pointlib/.

Datasets

KITTI

Prepare the dataset according to the datalists (*.txt in datasets)

datasets
|----kitti 
    |----depth_selection 
        |----val_selection_cropped
            |----...
        |----test_depth_completion_anonymous   
            |----...     
    |----rgb     
        |----2011_09_26
        |----...  
    |----train  
        |----2011_09_26_drive_0001_sync
        |----...      

Training

We will release the training code after the peer-review process.

Test

run

bash run_eval.sh

Note that, currently we only release the pretrained model with 32 channels.

Contact

Shanshan Zhao: [e-mail] or [e-mail]