Gaussian splatting
PulseAugur coverage of Gaussian splatting — every cluster mentioning Gaussian splatting across labs, papers, and developer communities, ranked by signal.
3 天有情绪数据
-
新方法使用真实视频训练人工智能物理模型
研究人员开发了一种新方法，可以直接从无标签的真实视频中训练神经对象动力学模型，克服了合成数据的局限性。该框架使用基于粒子的动力学模型并与高斯飞溅相结合，以预测粒子位置和旋转随时间的变化。这种方法无需显式的粒子级状态标签即可从真实视频中学习，并包含一个约 500 个视频的新数据集，展示了各种对象交互。
-
EndoGSim使用MLLM进行物理感知的手术模拟
研究人员开发了EndoGSim，一个用于机器人辅助手术中动态内窥镜场景模拟的新框架。该系统使用多模态大语言模型（MLLMs）来指导高斯喷溅，实现可变形组织和手术工具的物理感知重建。该框架集成了可微分的物质点法来优化物理属性，旨在提高手术模拟的保真度和准确性。
-
UAV-assisted scan-to-simulation framework enhances landslide modeling
Researchers have developed a new framework for simulating landslides that combines drone-based imagery with physics-informed Gaussian splatting. This approach captures photorealistic scenes and integrates them with the …
-
HumanSplatHMR refines 3D human pose and avatar generation via joint optimization
Researchers have introduced HumanSplatHMR, a novel framework that jointly optimizes 3D human pose estimation and avatar creation from video. This approach addresses limitations in existing methods by integrating pose re…
-
Diffusion models can now harmonize 3D Gaussian Splatting scenes
Researchers have developed a new method, D3DR, that leverages diffusion models to seamlessly insert 3D objects into existing 3D scenes rendered with Gaussian Splatting. The technique utilizes the diffusion model's impli…
-
New methods enable real-time global illumination and high-speed volumetric scene reconstruction
Researchers have developed a novel method for capturing and reconstructing high-speed dynamic 3D scenes using only standard, low-speed cameras. This technique encodes temporal information by illuminating the scene with …
-
Gaussian Splatting advances enable faster, more accurate wireless RF reconstruction
Two new research papers introduce Gaussian Splatting techniques adapted for wireless radiance field reconstruction. The first, BiSplat-WRF, proposes a planar Gaussian framework that incorporates electromagnetic coupling…
-
New SDDF representation offers faster, more accurate 3D scene reconstruction
Researchers have introduced a new 3D vision representation called the signed directional distance function (SDDF), designed to improve both reconstruction fidelity and rendering efficiency. Unlike existing methods like …
-
SketchFaceGS enables real-time 3D Gaussian head model editing from 2D sketches
Researchers have developed SketchFaceGS, a novel framework that enables real-time generation and editing of 3D Gaussian head models from 2D sketches. This system addresses the challenge of inferring detailed 3D structur…