发表机构
SuSTech; Shanghai AI Laboratory; HITSZ; Pengcheng Laboratory（南方科技大学; 上海人工智能实验室; 哈尔滨工业大学深圳校区; 鹏城实验室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究提出跨场景三维语义占用预测新任务，构建OccAnyScene框架实现室内外场景统一三维占用预测，在Occ-ScanNet和SurroundOcc-nuScenes数据集上取得最优mIoU结果。
AI 中文摘要
三维占用预测是场景理解的基础，但现有三维语义占用方法通常针对固定的场景类型和占用协议设计。我们提出跨场景三维语义占用预测这一新任务设定，要求单个模型处理异构的室内和室外场景，这些场景在相机、空间范围、体素规格及语义分类体系上存在差异。该设定带来了根本性挑战：在不同相机配置和场景尺度下实现度量一致且场景自适应的图像到三维提升。为应对这一挑战，我们提出OccAnyScene，这是一个基于预训练深度基础模型构建的以像素视锥体为中心的高斯框架。具体而言，该框架采用像素对齐视锥体特征聚合，为每个特征像素构建相机感知的视锥体查询，并采用视锥体参数化高斯构造，将每个查询解码为多个高斯，其位置和大小受预测的像素深度及对应视锥体几何约束。OccAnyScene取得了新的 state-of-the-art 结果，在室内 Occ-ScanNet 上达到59.92% mIoU，在室外 SurroundOcc-nuScenes 上达到23.06% mIoU。
英文摘要
3D occupancy prediction is fundamental to scene understanding, yet existing 3D semantic occupancy methods are typically specialized to fixed scene types and occupancy protocols. We introduce Cross-Scene 3D Semantic Occupancy Prediction, a new task setting which requires a single model to handle heterogeneous indoor and outdoor scenes with varying cameras, spatial ranges, voxel specifications, and semantic taxonomies. This setting poses a fundamental challenge: achieving metric-consistent yet scene-adaptive image-to-3D lifting across varying camera configurations and scene scales. To address this challenge, we propose OccAnyScene, a pixel-frustum-centered Gaussian framework built upon a pretrained depth foundation model. Specifically, the framework employs Pixel-Aligned Frustum Feature Aggregation to construct a camera-aware frustum query for each feature pixel, and Frustum-Parameterized Gaussian Construction to decode each query into multiple Gaussians whose positions and sizes are constrained by the predicted pixel depth and corresponding frustum geometry. OccAnyScene sets new state-of-the-art results, achieving 59.92% mIoU on the indoor Occ-ScanNet and 23.06% mIoU on the outdoor SurroundOcc-nuScenes.
CommentsFixed HTML display issue; manuscript unchanged
