BP-Net

Introduction

This is the pytorch implementation of our paper Bilateral Propagation Network for Depth Completion.

Environment

You can directly build the environment by running the following command if you use conda as the environment management tool.

conda env create -f environment.yml

Then you should have an env named bp.

Setup

Compile the C++ and CUDA code:

cd exts
python setup.py install

Dataset

We train and evaluate on KITTI and NYUV2 dataset.

KITTI

Please download KITTI depth completion dataset. The structure of data directory:

└── datas
    └── kitti
        ├── data_depth_annotated
        │   ├── train
        │   └── val
        ├── data_depth_velodyne
        │   ├── train
        │   └── val
        ├── raw
        │   ├── 2011_09_26
        │   ├── 2011_09_28
        │   ├── 2011_09_29
        │   ├── 2011_09_30
        │   └── 2011_10_03
        ├── test_depth_completion_anonymous
        │   ├── image
        │   ├── intrinsics
        │   └── velodyne_raw
        └── val_selection_cropped
            ├── groundtruth_depth
            ├── image
            ├── intrinsics
            └── velodyne_raw

NYUV2

We used preprocessed NYUv2 HDF5 dataset provided by Fangchang Ma. Note, the original full NYUv2 dataset is available at the official website.

The structure of data directory:

└── datas
    └── nyudepthv2
        ├── train
        │    ├── basement_0001a
        │    │    ├── 00001.h5
        │    │    └── ...
        │    ├── basement_0001b
        │    │    ├── 00001.h5
        │    │    └── ...
        │    └── ...
        └── val
            └── official
                ├── 00001.h5
                └── ...

Trained Models

You can directly download the trained models and extract them in checkpoints directory. Our models are trained on a GPU workstation with 4 Nvidia GTX 3090.

• BP_KITTI
• BP_NYU

Train

You can also train by yourself.

KITTI

train on KITTI

torchrun --nproc_per_node=4 --master_port 4321 train.py \
gpus=[0,1,2,3] num_workers=4 name=BP_KITTI \
net=PMP data=KITTI \
lr=1e-3 train_batch_size=2 test_batch_size=2 \
sched/lr=NoiseOneCycleCosMo sched.lr.policy.max_momentum=0.90 \
nepoch=30 test_epoch=25 ++net.sbn=true

NYUV2

train on NYUV2

torchrun --nproc_per_node=2 --master_port 1100 train.py \
gpus=[0,1] num_workers=4 name=BP_NYU \
net=PMP data=NYU data.num_mask=1 \
lr=2e-3 train_batch_size=8 test_batch_size=1 \
nepoch=100 test_epoch=80 ++net.sbn=true

Test

With the trained model, you can test and save results.

KITTI

test on KITTI selval set

python test.py gpus=[0] name=BP_KITTI ++chpt=BP_KITTI \
net=PMP num_workers=4 \
data=KITTI data.testset.mode=selval \
test_batch_size=1 metric=RMSE ++net.compile=true

test on KITTI test set and save for submission

python test.py gpus=[0] name=BP_KITTI ++chpt=BP_KITTI \
net=PMP num_workers=4 \
data=KITTI data.testset.mode=test data.testset.height=352 \
test_batch_size=1 metric=RMSE ++save=true

NYUV2

test on NYUV2 test set

python test.py gpus=[0] name=BP_NYU ++chpt=BP_NYU \
net=PMP num_workers=4 \
data=NYU test_batch_size=1 metric=MetricALL

Citation

If you find this work useful in your research, please consider citing:

@article{BP-Net,
title={Bilateral Propagation Network for Depth Completion},
author={Tang, Jie and Tian, Fei-Peng and An, Boshi and Li, Jian and Tan, Ping},
journal={CVPR},
year={2024}
}