Tag
Skyfall AI unveiled SKY-001, an engineering world model that reconstructs a Jaguar car with thousands of 3D parts, claiming CAD accuracy on par with GPT-6 Astra while running 51× faster and costing 150× less.
HiP3R 提出一种层级式位姿图方法，通过先优化块级 Sim(3) 图建立全局布局，再固定该布局估计每帧共享的 SE(3) 相机位姿，将局部视频重建合并为全局一致的相机轨迹和 3D 几何，无需重新训练。
LEGO-SCD is an online scene change detection method that learns scene-specific statistical models from automatic edits to a reference 3D Gaussian Splatting scene without manual annotations, enabling closed-form Bayesian evidence accumulation with object-guided mask refinement; it achieves 67.64% mIoU on PASLCD and 75.82% on CCS3D at 17.2 FPS.
EgoXperience 提出一个统一框架，通过自动标注的大规模数据和单次前向传播模型，从第一人称视频中联合重建相机运动、场景几何、手部与物体，并发布了 EgoX-Data-6M 数据集与 EgoX-Bench 基准，取得显著优于基线的性能。
展示了 Gemini 4 Argon 结合 Google Antigravity 和 Blender MCP，将碎片化的考古数据重建为可探索的 3D 古迹世界，助力数字文化遗产保护。
Turing Post 深度科普指南：讲解 3D Gaussian Splatting 如何作为具身智能机器人的空间记忆，把拍摄场景重建为可搜索、可更新的 3D 表示，并延伸到物理仿真理解。
The paper introduces Seed Injection and Temporal Biased Relocation to allocate Gaussian primitives based on predicted spatiotemporal destinations rather than current locations, improving reconstruction of small fast-moving objects and fine textures in dynamic scenes without increasing the Gaussian budget.
InfiniSplat v2 通过解耦几何与属性预测、使用真实相机位姿对齐以及大角度视角监督，显著提升前馈式 3D 高斯重建的新视角渲染质量，在六个数据集上取得 SOTA 的 LPIPS 和 DISTS 指标。
WorldRecon 提出一种生成式多视角网格重建框架，可从无位姿 RGB 图像重建大规模 3D 场景的完整网格。它将 TRELLIS.2 生成模型与重建式条件架构结合，通过邻域感知潜在修补和因果世界生成保证场景级一致性，在合成与真实室内场景上分别将 Chamfer 距离降低 14.9% 和 43.1%，达到最先进性能。
CineRecon introduces a benchmark with 50 reference videos and 360 annotated evaluation clips to test frontier models' ability to reconstruct editable dynamic 3D scenes from videos, evaluating 12 models across camera, static, and dynamic consistency dimensions and finding failures in all three.
UniForge3D 提出在共享的锚点引导空间中联合建模 3D 重建与生成，解决相机相对重建与规范空间生成之间的坐标不匹配问题，并在 Toys4K 和 GSO 数据集上取得了最先进的外观与几何结果。
Drive3R 提出一种块状流式几何基础模型，通过块状因果注意力和置信度感知重锚定实现可扩展的长时程多视图 3D 重建，在 nuScenes、KITTI 和 Waymo 上达到新的最先进水平。
NVIDIA researchers present GATOR, a generative and agentic framework that reconstructs complete, textured 3D object assets and their scene-relative pose from one or more casual images, using a local modality mixer, text-guided semantic conditioning, and an edit-render-review agent loop for refinement.
JoHan：一个从视频序列直接联合生成2D和3D手部姿态序列的统一生成框架，无需依赖逐帧姿态预测，在遮挡和不完整观测下显著提升手部运动重建的精度、速度和平滑度。
CADFather is an autonomous agentic system that reconstructs editable parametric CAD programs from 3D meshes by coordinating a vision-language assistant with complementary learned, algorithmic, and optimization tools, evaluated on benchmarks like DeepCAD, Fusion360, and CADENA-Bench without additional training.
提出了MEGA（Mesh Extraction from Gaussians）这一“先分割后网格化”框架，通过Spatial Visual Distillation（SVD）和掩码引导的神经曲面重建，从3DGS场景中提取物体级、水密的网格，实现高质量的物体级几何占用和物理交互。
T3lescope提出一种单一固定分辨率生成器的推理时粗到细级联方法，从多视角图像实现跨尺度（从单物体到大型户外场景）的任意分辨率高保真3D场景网格重建，无需逐场景优化，在室内、户外和城市规模场景上超越了前馈与生成基线。
DyRAD presents a radar novel view synthesis method for dynamic driving scenes that models static background and motion-tracked dynamic point reflectors to render complete range-azimuth-Doppler tensors, using a fixed analytic point-spread function to enable zero-shot sensor-configuration transfer and outperforming baselines by a wide margin on the RADIal benchmark.
GPT-6 Astra introduces the ability to convert real-room videos into interactive 3D environments, enabling robot training and novel viewpoint rendering for AI video models.
FurE is an efficient strand-based method for 3D animal fur reconstruction from multi-view images, using a PCA decoder learned from human-hair data to overcome dataset scarcity and achieve 10x speedup over current methods.