/shadow_pose_estimation

pose estimation with shadow image enhancement

Primary LanguagePureBasicMIT LicenseMIT

shadow_pose_estimation

pose estimation with shadow image enhancement

Usage

  • download the OTS dataset and put the dataset in your desired path.
  • set up the image path in train.py (specifically the im_path, img_B path)
  • setup the image path in data_load.py
  • setup the save path for model
  • train image enhancement model model by running command
cd image_enhancement
python3 train.py
  • put the test image path in your desired path
  • modifiy the path for test images and saved model in test.py
  • setup the saved path for the enhanced images
  • run test.py py to run the inference results for image enhancement
  • if you need to run pose estimation results on enhanced images, please run the command
cd pose_estimation
python3 get_pose.py --image_dir=SAVED_ENHANCED_PATH

Image Enhancement Network Design

Overall arch

EM arch

System setup

Acknowledgements