Supplementary code and training data for the wide FoV person segmentation approach.
The code will and training data will be made publicly available on github on acceptance and serves as reference for the reviewing process.
The training code is seperated into seperated files for the 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.