@misc{indiciaef9f64e65e777,
  title = {MOBIUS: Big-to-Mobile Universal Instance Segmentation via Multi-modal Bottleneck Fusion and Calibrated Decoder Pruning},
  author = {Mattia Segu and Marta Tintore Gazulla and Yongqin Xian and Luc Van Gool and Federico Tombari},
  year = {2025},
  url = {https://arxiv.org/abs/2510.15026},
  note = {Source identifier: 2510.15026}
}