双分辨率编码器提升3D占位预测精度，兼顾细节与全局上下文。
DualPathOcc: Dual-Resolution BEV Encoder for 3D Occupancy Prediction
- 先用高分辨特征聚合再压缩到鸟瞰图，保留几何细节
- 37.37的mIoU在Occ3D-nuScenes上领先，无需显式深度监督
- 适合关注自动驾驶中场景理解的开发者
从多视角图像预测3D占位，需在2D转3D过程中保持几何细节，并对稀疏体素场景进行推理。我们提出DualPathOcc，一种基于相机的框架：包含空间增强模块，在鸟瞰图压缩前进行高分辨率特征聚合；采用SENet增强的双路径鸟瞰图编码器，实现局部-全局上下文建模；以及高度感知加权交叉熵损失，优化近地占位预测。最终模型仅使用占位监督，无显式深度损失。在单帧Occ3D-nuScenes数据集上，达到37.37 mIoU。进一步分析了以表面为中心的深度目标如何与体素占位学习相互作用。
原文摘要 · Abstract (English)
Predicting 3D occupancy from multi-view images requires preserving geometric detail during 2D-to-3D lifting while reasoning over sparse, volumetric scene representations. We present DualPathOcc, a camera-based framework that combines a Spatial Enhancer for high-resolution feature aggregation before BEV compression, a SENet-augmented dual-path BEV encoder for local-global context modeling, and height-aware weighted cross-entropy for near-ground occupancy. The final model is optimized with occupancy supervision and no explicit depth loss. On single-frame Occ3D-nuScenes, DualPathOcc achieves 37.37 mIoU. We further analyze how surface-centered depth targets interact with volumetric occupancy learning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。