AV-GS

Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis

Homepage Paper

AV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis
Swapnil Bhosale, Haosen Yang, Diptesh Kanojia, Jiangkang Deng, Xiatian Zhu
NeurIPS 2024

This repository is the official implementation of "AV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis". In this paper, we propose a material-aware and geometry-aware condition for binaural audio synthesis, by learning an explicit point-based 3D audio-visual scene representation with an audio-guidance parameter on locally initialized Gaussian points, taking into account the space relation from the listener and sound source.

Updates

• Code coming soon!

🛠️ Pipeline

📜 BibTex

@article{bhosale2024avgs,
  title={AV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis},
  author={Swapnil Bhosale, Haosen Yang, Diptesh Kanojia, Jiangkang Deng, Xiatian Zhu},
  journal={Advances in Neural Information Processing Systems},
  year={2024}
}