@misc{indiciaee938d8f00d88,
  title = {Decoupled Visual Processing: Efficient Multimodal Adaptation via Modality-Specific Transformer Substitution},
  author = {Mingkuan Feng and Zhengqi Wen and Jianhua Tao},
  year = {2026},
  url = {https://arxiv.org/abs/2607.26596},
  note = {Source identifier: 2607.26596}
}