Tag
Velox 提出一种免训练的注意力路由器，利用相机视锥切片的几何对应关系进行区域级上下文选择，从而加速长视频生成中的 KV 缓存处理。在第 48 个 chunk 时注意力加速 4.36 倍、去噪加速 2.43 倍，且在 WBench 导航基准上几乎不损失质量。
OmniCam 提出一种自回归模型，可从全景图与文本描述生成相机位姿序列，通过几何锚定的姿态 token 学习在 OmniCaT 数据集上将轨迹误差降低 28–47%、碰撞率降低 65.8%，并应用于相机可控视频生成与机器人主动感知。
The user successfully used edulab Skills to generate 3D geometry problem-solving demonstrations on the Qwen3.8 27B model and discussed performance and optimization needs under specific hardware configurations.
This paper introduces robot-centric pointmaps, which encode 3D scene coordinates in the robot frame directly into image pixels, to resolve the frame mismatch between camera observations and robot action definitions in Vision-Language-Action models. The method improves Pi0.5 and SmolVLA on RoboCasa and generalizes better to unseen camera placements in real-robot experiments.
World Tracing introduces a generative pixel-aligned geometry representation that predicts 3D points aligned with observed pixels while completing occluded surfaces. It uses a diffusion transformer trained with pixel-space flow matching, achieving strong performance on visible-surface reconstruction and complete geometry generation across object, scene, and dynamic benchmarks.
ViGeo is a transformer-based foundation model that recovers dense and consistent 3D geometry from videos using dynamic chunking attention and a completion-based data refinement framework, achieving state-of-the-art performance across multiple tasks.