Supplementary code and training data for the wide FoV person segmentation approach. The paper can be found here.
The training code is seperated for different tasks:
- Person Detection -->
person_detection.ipynb
- Action Detection -->
action_detection.ipynb
- Inference code -->
test_tiago.ipynb
The dataset are contained in separate sub-folders inside the images
folder depending on the corresponding experiment set. The data is provided with annotations in the COCO format.