Human Detection and Segmentation via Multi-view Consensus

This is the source code for the ICCV 2021 paper "Human Detection and Segmentation via Multi-view Consensus".

Dependencies

Python 3.6

PyTorch 1.5

Cuda 9.2

Dataset

Please contact the authors (isinsu.katircioglu@gmail.com, joerg.spoerri@balgrist.ch) to have access to the Ski-PTZ dataset.

Training

• Download the Ski-PTZ dataset. The dataset should include subfolders named Videos_small, Videos, Data and Segmentation_masks.
• Change the dataset path in datasets/dataset_factory.py.
• Download the pre-trained models.
• Create a folder named pretrained in the cloned directory and put the pre-trained models inside.

You can start training with the following command:

$ cd src
$ python configs/run_config.py configs/config_dict_mvc_ski.py

Citation

Please cite the paper in your publications if it helps your research:

@inproceedings{Katircioglu21,
  Author = {I. Katircioglu and H. Rhodin and J. Sp{\"o}rri and M. Salzmann and P. Fua},
  Booktitle = ICCV,
  Title = {{Human Detection and Segmentation via Multi-view Consensus}},
  Year = 2021}