Look Who's Talking: Active Speaker Detection in the Wild

Dependencies

pip install -r requirements.txt

In addition to the Python dependencies, ffmpeg must be installed on the system.

Instructions

First, download the videos to $DATA_DIR/original.

Run the following to convert the videos and visualise the labels.

python3 run_convert.py --data_dir $DATA_DIR
python3 run_visualize.py --data_dir $DATA_DIR

Citation

Please cite the following if you make use of the code.

@inproceedings{kim2021you,
  title={Look Who's Talking: Active Speaker Detection in the Wild},
  author={Kim, You Jin and Heo, Hee-Soo Heo and Choe, Soyeon and Chung, Soo-Whan and Kwon, Yoohwan and Lee, Bong-Jin and Kwon, Youngki and Chung, Joon Son},
  booktitle={Interspeech},
  year={2021}
}

GitHub

https://github.com/clovaai/lookwhostalking