• Contributions

  • We build a dataset to evaluate the performance of existing MLLMs on 360° image perception tasks.

  • We develop a method to improve MLLM's abilities on these tasks. 

• Code and Dataset

  • Our code and dataset will be publicly released at https://github.com/TranHuyen1191/360Bench-Free360

• Acknowledgments

  • If you use the code and dataset, please cite our paper.

@misc{tran2026360degimageperceptionmllms,

      title={360{\deg} Image Perception with MLLMs: A Comprehensive Benchmark and a Training-Free Method}, 

      author={Huyen T. T. Tran and Van-Quang Nguyen and Farros Alferro and Kang-Jun Liu and Takayuki Okatani},

      year={2026},

      eprint={2603.16179},

      archivePrefix={arXiv},

      primaryClass={cs.CV},

      url={https://arxiv.org/abs/2603.16179}, 

}