1 Fast-Convolutional-for-Multi-Channel-Objecet-Detection

Fast convolutional for multi-channel object detection and evaluation inspired by SFA3D

SFA3D: " Super Fast and Accurate 3D Object Detection based on 3D LiDAR Point Clouds" ''' @misc{Super-Fast-Accurate-3D-Object-Detection-PyTorch, author = {Nguyen Mau Dung}, title = {{Super-Fast-Accurate-3D-Object-Detection-PyTorch}}, howpublished = {\url{https://github.com/maudzung/Super-Fast-Accurate-3D-Object-Detection}}, year = {2020} } '''

2. DATASET

3D KITTI Dataset은Link. 에서 다운 받을 수 있습니다. 구성 요소는 다음과 같습니다:

• Velodyne point clouds (29 GB)
• Training labels of object data set (5 MB)
• Camera calibration matrices of object data set (16 MB)
• Left color images of object data set (12 GB) (For visualization purpose only)

3. Train

Train Code

''' !python train.py --gpu_idx 0 --mode classic #--mode checkpoints folder '''

4. Evaluate (AP)

Evaluate Code

''' !python train.py --evaluate --gpu_idx 0 --resume_path './model_path' '''

5. Visualization (Demo)

Demo Code

''' !python demo_2_sides.py --gpu_idx 0 --peak_thresh 0.2 --saved_fn final_demo_classic '''

6. Hierarchy

${ROOT}
└── checkpoints/
    ├── fpn_resnet_18/    
        ├── fpn_resnet_18_epoch_300.pth
└── dataset/    
    └── kitti/
        ├──ImageSets/
        │   ├── test.txt
        │   ├── train.txt
        │   └── val.txt
        ├── training/
        │   ├── image_2/ (left color camera)
        │   ├── calib/
        │   ├── label_2/
        │   └── velodyne/
        └── testing/  
        │   ├── image_2/ (left color camera)
        │   ├── calib/
        │   └── velodyne/
        └── classes_names.txt
└── sfa/
    ├── config/
    │   ├── train_config.py
    │   └── kitti_config.py
    ├── data_process/
    │   ├── kitti_dataloader.py
    │   ├── demo_dataset.py
    │   ├── kitti_bev_utils.py
    │   ├── kitti_dataset.py
    │   ├── transformation.py
    │   └── kitti_data_utils.py
    ├── losses/
    │   └── losses.py
    ├── models/
    │   ├── fpn_resnet.py
    │   ├── resnet.py
    │   └── model_utils.py
    └── utils/
    │   ├── box_np_ops.py
    │   ├── classic_utils.py
    │   ├── demo_utils.py
    │   ├── eval.py
    │   ├── evaluate.py
    │   ├── evaluation_utils.py
    │   ├── kitti_common.py
    │   ├── logger.py
    │   ├── lr_scheduler.py
    │   ├── misc.py
    │   ├── nms_gpu.py
    │   ├── rotate_iou.py
    │   ├── nms_gpu.py
    │   ├── rotate_iou.py
    │   ├── torch_utils.py
    │   ├── train_utils.py
    │   └── visualization_utils.py
    ├── classical_demo.ipynb
    ├── classical_train_eval.ipynb
    ├── demo_2_sides.py
    ├── demo_front.py
    ├── test.py
    └── train.py
├── README.md 
└── requirements.txt

REFERENCE

[1] "Object Detector for Autonomous Vehicles Based on Improved Faster RCNN": 2D 이전 버전
[2]"Torch-quantum"QNN Implementation
[3] "KITTI-WAYMO Adapter": WAYMO데이터 활용
[4] CenterNet: Objects as Points paper, PyTorch Implementation
[5] RTM3D: PyTorch Implementation
[6] Libra_R-CNN: PyTorch Implementation

The YOLO-based models with the same BEV maps input:
[7] Complex-YOLO: v4, v3, v2

3D LiDAR Point pre-processing:
[8] VoxelNet: PyTorch Implementation