@misc{indiciaed1a00453a573,
  title = {Towards Bitstream-corrupted Harsh Visual Understanding: Through Bitstream Language Modeling as Robust Semantic Priors},
  author = {Chaoran Huang and Fangcheng Li and Tianyi Liu and Wenyang Liu and Kejun Wu},
  year = {2026},
  url = {https://arxiv.org/abs/2608.21837},
  note = {Source identifier: 2608.21837}
}