@misc{indiciae8b6396881b8a,
  title = {Segmentation before Answering: Pixel Grounding for MLLM Visual Reasoning},
  author = {Yake Wei and Yuan Wang and Fengyun Rao and Jing Lyu and Di Hu},
  year = {2026},
  url = {https://arxiv.org/abs/2607.05798},
  note = {Source identifier: 2607.05798}
}