zjc
/
DMPR-PS
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
车位角点检测代码
9 Commits 1 Branch 0 Tags 569 KiB
Python 100%
Go to file
Teoge 96ba10819c
Create LICENSE 		2019-06-03 10:21:00 +08:00
data Release 2019-04-07 14:11:38 +08:00
model Release 2019-04-07 14:11:38 +08:00
util Release 2019-04-07 14:11:38 +08:00
.gitignore Initial Commit 2018-07-20 16:25:15 +08:00
LICENSE Create LICENSE 2019-06-03 10:21:00 +08:00
README.md Add README 2019-05-23 17:03:20 +08:00
collect_thresholds.py Refine thresholds 2018-11-21 15:45:22 +08:00
config.py Add README 2019-05-23 17:03:20 +08:00
evaluate.py Release 2019-04-07 14:11:38 +08:00
inference.py Release 2019-04-07 14:11:38 +08:00
prepare_dataset.py Refine thresholds 2018-11-21 15:45:22 +08:00
ps_evaluate.py Release 2019-04-07 14:11:38 +08:00
train.py Release 2019-04-07 14:11:38 +08:00

README.md

DMPR-PS

This is the implementation of DMPR-PS using PyTorch.

Requirements

  • CUDA
  • PyTorch
  • OpenCV
  • NumPy
  • Pillow
  • Visdom (optional)
  • Matplotlib (optional)

Pre-trained weights

The pre-trained weights could be used to reproduce the number in the paper.

Inference

  • Image inference

    python inference.py --mode image --detector_weights $DETECTOR_WEIGHTS

  • Video inference

    python inference.py --mode video --detector_weights $DETECTOR_WEIGHTS --video $VIDEO

    DETECTOR_WEIGHTS is the trained weights of detector.
    VIDEO is path to the video.
    View config.py for more argument details.

Prepare data

  1. Download ps2.0 from here, and extract.

  2. Download the labels, and extract.

  3. Perform data preparation and augmentation:

    python prepare_dataset.py --dataset trainval --label_directory $LABEL_DIRECTORY --image_directory $IMAGE_DIRECTORY --output_directory $OUTPUT_DIRECTORY
python prepare_dataset.py --dataset test --label_directory $LABEL_DIRECTORY --image_directory $IMAGE_DIRECTORY --output_directory $OUTPUT_DIRECTORY

    LABEL_DIRECTORY is the directory containing json labels.
    IMAGE_DIRECTORY is the directory containing jpg images.
    OUTPUT_DIRECTORY is the directory where output images and labels are.
    View prepare_dataset.py for more argument details.

Train

python train.py --dataset_directory $TRAIN_DIRECTORY

TRAIN_DIRECTORY is the train directory generated in data preparation.
View config.py for more argument details (batch size, learning rate, etc).

Evaluate

  • Evaluate directional marking-point detection

    python evaluate.py --dataset_directory $TEST_DIRECTORY --detector_weights $DETECTOR_WEIGHTS

    TEST_DIRECTORY is the test directory generated in data preparation.
    DETECTOR_WEIGHTS is the trained weights of detector.
    View config.py for more argument details (batch size, learning rate, etc).

  • Evaluate parking-slot detection

    python evaluate.py --label_directory $LABEL_DIRECTORY --image_directory $IMAGE_DIRECTORY --detector_weights $DETECTOR_WEIGHTS

    LABEL_DIRECTORY is the directory containing testing json labels.
    IMAGE_DIRECTORY is the directory containing testing jpg images.
    DETECTOR_WEIGHTS is the trained weights of detector.
    View config.py for more argument details.