提出无需深度信息的3D旋转增强方法，提升单目3D检测精度
3DRot: Rediscovering the Missing Primitive for RGB-Based 3D Augmentation
- 基于相机光心旋转图像并同步更新内外参与标注，保持几何一致性
- 在SUN RGB-D上使3D IoU提升至44.51，旋转误差降低至20.93°
- 适用于单目/多模态3D任务，可无缝集成现有模型训练流程
基于RGB的3D任务如3D检测、深度估计等仍面临标注稀缺、数据增强工具匮乏的问题，因多数图像变换（如旋转、扭曲）会破坏几何一致性。尽管水平翻转和色彩抖动已是常规操作，但严谨的3D旋转增强却长期缺失，主要源于误认为其需依赖场景深度或重建。本文提出3DRot，一种即插即用的增强方法，通过围绕相机光心旋转并同步更新RGB图像、相机内参、物体姿态及3D标注，实现无深度依赖的几何一致旋转与镜像。在单目3D检测任务中，将3DRot加入冻结的DINO-X + Cube R-CNN框架，在SUN RGB-D上使$IoU_{3D}$从43.21提升至44.51，旋转误差（ROT）从22.91°降至20.93°，$mAP_{0.5}$从35.70升至38.11；跨域IN10数据集亦见稳定增益。在深度估计任务中，结合标准BTS增强策略，使NYU Depth v2的abs-rel从0.1783降至0.1685，$δ<1.25$从0.7472升至0.7548，并降低跨数据集误差。在KITTI上，于MVX-Net（LiDAR+RGB）中应用相同旋转策略，3D AP由约63.85提升至65.16，且兼容标准3D增强。
原文摘要 · Abstract (English)
RGB-based 3D tasks, e.g., 3D detection, depth estimation, 3D keypoint estimation, still suffer from scarce, expensive annotations and a thin augmentation toolbox, since many image transforms, including rotations and warps, disrupt geometric consistency. While horizontal flipping and color jitter are standard, rigorous 3D rotation augmentation has surprisingly remained absent from RGB-based pipelines, largely due to the misconception that it requires scene depth or scene reconstruction. In this paper, we introduce 3DRot, a plug-and-play augmentation that rotates and mirrors images about the camera's optical center while synchronously updating RGB images, camera intrinsics, object poses, and 3D annotations to preserve projective geometry, achieving geometry-consistent rotations and reflections without relying on any scene depth. We first validate 3DRot on a classical RGB-based 3D task, monocular 3D detection. On SUN RGB-D, inserting 3DRot into a frozen DINO-X + Cube R-CNN pipeline raises $IoU_{3D}$ from 43.21 to 44.51, cuts rotation error (ROT) from 22.91$^\circ$ to 20.93$^\circ$, and boosts $mAP_{0.5}$ from 35.70 to 38.11; smaller but consistent gains appear on a cross-domain IN10 split. Beyond monocular detection, adding 3DRot on top of the standard BTS augmentation schedule further improves NYU Depth v2 from 0.1783 to 0.1685 in abs-rel (and 0.7472 to 0.7548 in $δ<1.25$), and reduces cross-dataset error on SUN RGB-D. On KITTI, applying the same camera-centric rotations in MVX-Net (LiDAR+RGB) raises moderate 3D AP from about 63.85 to 65.16 while remaining compatible with standard 3D augmentations.
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。