@misc{indiciaefb5e182e9940,
  title = {iGVLM: Dynamic Instruction-Guided Vision Encoding for Question-Aware Multimodal Understanding},
  author = {Hanpeng Liu and Yaqian Li and Zidan Wang and Shuoxi Zhang and Zihao Bo and Rinyoichi Takezoe and Kaiwen Long and Kun He},
  year = {2026},
  url = {https://arxiv.org/abs/2603.02748},
  note = {Source identifier: 2603.02748}
}