P2VANet

Public-Private Attributes-Based Variational Adversarial Network for Audio-Visual Cross-Modal Matching

requirments

python 3.8 librosa 0.7.2 numpy 1.19.0 torch 1.4.0 torchvision 0.5.0

Get Dataset and Paper

• Download the VoxCeleb, VGGFace
• Latest Paper List Audio-visual matching

VoxCeleb1

• wav audio data, 1,251 people in total, 39 GB after decompression.

• Baidu Cloud link: VoxCeleb1

• Decompression command:

• zip -s 0 split.zip --out unsplit.zip

• unzip unslit.zip

• Vox1 official website: VoxCeleb1

VoxCeleb2

• MP4 video data, files include audio, total of 5,994 people, 255 GB after decompression.

• Baidu Cloud link: VoxCeleb2

• Decompression command:

• zip -s 0 vox2_mp4_dev.zip --out unsplit.zip

• unzip unslit.zip

• Vox2 official website: VoxCeleb2

Contact

If you have any questions, please feel free to contact with me at Netizenwjx@foxmail.com.

Citation

@Article{Zheng2024Public,
  author={Zheng, Aihua and Yuan, Fan and Zhang, Haichuan and Wang, Jiaxiang and Tang, Chao and Li, Chenglong},
  journal={IEEE Transactions on Circuits and Systems for Video Technology}, 
  title={Public-Private Attributes-Based Variational Adversarial Network for Audio-Visual Cross-Modal Matching}, 
  volume={34},
  number={9},
  pages={8698-8709},
  note = {doi: 10.1109/TCSVT.2024.3390573},
  year={2024}}