PanoWorld: Towards Spatial Supersensing in 360^circ Panorama World

Changpeng Wang, Xin Lin, Junhan Liu, Yuheng Liu, Zhen Wang, Donglian Qi, Yunfeng Yan, Xi Chen

arXiv:2605.13169 · 2026-05-15 공개 · arXiv · PDF

instruction-tuning spatial-reasoning cross-attention spherical-geometry depth-aware pano-native erp-panorama panospace-bench

Abstract

Multimodal large laboratory models (MLLMs) still struggle with spatial understanding under the dominant perspective-image paradigm, which inherits the narrow field of view of human-like perception. For navigation, robotic search, and 3D scene understanding, 360-degree panoramic sensing offers a form of supersensing by capturing the entire surrounding environment at once. However, existing MLLM pipelines typically decompose panoramas into multiple perspective views, leaving the spherical structure of equirectangular projection (ERP) largely implicit. In this paper, we study pano-native understanding, which requires an MLLM to reason over an ERP panorama as a continuous, observer-centered space. To this end, we first define the key abilities for pano-native understanding, including semantic anchoring, spherical localization, reference-frame transformation, and depth-aware 3D spatial reasoning. We then build a large-scale metadata construction pipeline that converts mixed-source ERP panoramas into geometry-aware, language-grounded, and depth-aware supervision, and instantiate these signals as capability-aligned instruction tuning data. On the model side, we introduce PanoWorld with Spherical Spatial Cross-Attention, which injects spherical geometry into the visual stream. We further construct PanoSpace-Bench, a diagnostic benchmark for evaluating ERP-native spatial reasoning. Experiments show that PanoWorld substantially outperforms both proprietary and open-source baselines on PanoSpace-Bench, H* Bench, and R2R-CE Val-Unseen benchmarks. These results demonstrate that robust panoramic reasoning requires dedicated pano-native supervision and geometry-aware model adaptation. All source code and proposed data will be publicly released.

한국어 요약

📋 한 줄 요약

**[Multimodal/Spatial Reasoning]** 360° ERP 파노라마를 관측자 중심 연속 공간으로 직접 추론하는 pano-native MLLM PanoWorld, Spherical Spatial Cross-Attention으로 구면 기하를 시각 스트림에 주입.

🎯 핵심 기여도

💡 핵심 아이디어

360° 파노라마를 perspective view로 쪼개면 구면 구조와 관측자 중심성이 사라진다. PanoWorld는 ERP 표현을 분해하지 않고 그대로 받아, 구면 좌표 기하를 어텐션에 명시적으로 주입해 모델이 "주변 전체 공간"을 일관된 좌표계로 추론하게 한다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: MLLM의 공간 이해를 perspective 시야의 한계 너머로 확장, 내비게이션·로보틱 탐색·3D 씬 이해 같은 supersensing 응용의 새로운 표준 표현을 제시한다. **한계**: ERP 입력·구면 어텐션에 특화되어 일반 perspective 이미지·동영상에서는 별도 어댑테이션이 필요하며, 학습 데이터가 합성·시뮬레이션에 일부 의존한다.

🚀 실용적 활용