chensi521/dd3d_

不使用docker安装的老版本

dd3d-supplement

fork from https://github.com/TRI-ML/dd3d.git

DD3D: "Is Pseudo-Lidar needed for Monocular 3D Object detection?"

Official PyTorch implementation of DD3D: Is Pseudo-Lidar needed for Monocular 3D Object detection? (ICCV 2021),

Datasets

By default, datasets are assumed to be downloaded in /data/datasets/<dataset-name> (can be a symbolic link). The dataset root is configurable by DATASET_ROOT.

KITTI

The KITTI 3D dataset used in our experiments can be downloaded from the KITTI website. For convenience, we provide the standard splits used in 3DOP for training and evaluation:

The dataset must be organized as follows:

<DATASET_ROOT>
    └── KITTI3D
        ├── mv3d_kitti_splits
        │   ├── test.txt
        │   ├── train.txt
        │   ├── trainval.txt
        │   └── val.txt
        ├── testing
        │   ├── calib
        |   │   ├── 000000.txt
        |   │   ├── 000001.txt
        |   │   └── ...
        │   └── image_2
        │       ├── 000000.png
        │       ├── 000001.png
        │       └── ...
        └── training
            ├── calib
            │   ├── 000000.txt
            │   ├── 000001.txt
            │   └── ...
            ├── image_2
            │   ├── 000000.png
            │   ├── 000001.png
            │   └── ...
            └── label_2
                ├── 000000.txt
                ├── 000001.txt
                └── ..

Install

sudo apt install tk-dev
sudo apt install python3-tk

pip install -r requirements.txt

Pre-trained DD3D models

The DD3D models pre-trained on dense depth estimation using DDAD15M can be downloaded here:

backbone	download
DLA34	[model] (链接: https://pan.baidu.com/s/1vpncowhJjhivGNrqMLSTYg 提取码: 4f6q)

(Optional) Eigen-clean subset of KITTI raw.

To train our Pseudo-Lidar detector, we curated a new subset of KITTI (raw) dataset and use it to fine-tune its depth network. This subset can be downloaded here. Each row contains left and right image pairs. The KITTI raw dataset can be download here.

Validating installation

To validate and visualize the dataloader (including data augmentation), run the following:

python3 visualize_dataloader.py +experiments=dd3d_kitti_dla34 SOLVER.IMS_PER_BATCH=4

To validate the entire training loop (including evaluation and visualization), run the overfit experiment (trained on test set):

python3 train.py +experiments=dd3d_kitti_dla34

Predict

python3 predict.py +experiments=dd3d_kitti_dla34

Evaluation

One can run only evaluation using the pretrained models:

python3 train.py +experiments=dd3d_kitti_dla34 EVAL_ONLY=True MODEL.CKPT=<path-to-pretrained-model>
# use smaller batch size for single-gpu
python3 train.py +experiments=dd3d_kitti_dla34  EVAL_ONLY=True MODEL.CKPT=<path-to-pretrained-model> TEST.IMS_PER_BATCH=4

License

The source code is released under the MIT license. We note that some code in this repository is adapted from the following repositories:

• detectron2
• AdelaiDet

Reference

@inproceedings{park2021dd3d,
  author = {Dennis Park and Rares Ambrus and Vitor Guizilini and Jie Li and Adrien Gaidon},
  title = {Is Pseudo-Lidar needed for Monocular 3D Object detection?},
  booktitle = {IEEE/CVF International Conference on Computer Vision (ICCV)},
  primaryClass = {cs.CV},
  year = {2021},
}