3D Gaussians
PulseAugur coverage of 3D Gaussians — every cluster mentioning 3D Gaussians across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
DecoGS method improves 3D video streaming with adaptive updates
Researchers have developed DecoGS, a novel method for efficiently streaming 3D reconstructions from video. Unlike previous approaches that update every element in a 3D scene, DecoGS intelligently focuses optimization on…
-
New 3D Point Splatting Renderer Advances mmWave Radar Novel View Synthesis
Researchers have developed 3D Point Splatting (3DPS), a novel differentiable point renderer specifically designed for mmWave radar novel view synthesis. This new method addresses limitations in existing techniques by be…
-
New inverse rendering method uses line segments for fuzzy geometry
Researchers have developed a new inverse rendering method that reconstructs fuzzy geometry using explicit line segments, overcoming limitations of current volumetric primitive methods like 3D Gaussians. This approach ra…
-
VersaGauss framework enables realistic multiphase dynamics simulation with 3D Gaussians
Researchers have introduced VersaGauss, a novel framework designed for generating and simulating complex 3D dynamic scenes with multiphase interactions. This system takes a few images as input and produces realistic, ph…
-
AvatarDynamizer transforms static 3D avatars into dynamic 4D versions
Researchers have developed AvatarDynamizer, a novel generative method designed to transform static 3D avatars into dynamic, realistic, and multi-view-consistent 4D avatars. This approach tackles the limitations of exist…
-
SplatGuide uses 3D Gaussians for pose-free novel view synthesis · arXiv
Researchers have introduced SplatGuide, a novel method for generating photorealistic novel views from unposed images. This approach leverages geometric priors from 3D Gaussians, reusing a single 3D Gaussian Splatting (3…
-
New frameworks unify 3D scene understanding and generation for autonomous driving · 2 sources tracked
Researchers have developed two new frameworks, USR-Drive and GaussianDWM++, that unify 3D scene understanding and generation for autonomous driving. USR-Drive jointly denoises 3D Gaussian primitives and bounding boxes u…
-
New research advances 3D scene generation for AI and autonomous driving
Two recent arXiv papers explore advancements in 3D scene generation, a field crucial for applications like autonomous driving and virtual reality. The first paper, a survey, categorizes current methods into procedural, …
-
Struct-GStream enables efficient free-viewpoint video streaming with 3D Gaussians
Researchers have developed Struct-GStream, a novel method for efficient free-viewpoint video (FVV) streaming using structured 3D Gaussians. This approach addresses the challenges of real-time rendering and high storage …
-
New TOM-GS method enables editable video with temporal opacity modulation
Researchers have introduced TOM-GS, a novel method for creating editable video representations using static 3D Gaussians with temporal opacity modulation. This approach allows static spatial components to smoothly fade …
-
New method computes fast pointwise signed distance for point clouds
Researchers have developed a novel method for computing signed distances to point clouds, enabling fast, high-resolution evaluations of the approximate underlying surface. This technique involves locally fitting point c…
-
New SAGO framework enables real-time 3D Gaussian segmentation
Researchers have developed SAGO (Segment Any Gaussians Online), a new framework designed to enable real-time interactive segmentation of 3D Gaussian Splatting (3DGS) scenes. Unlike previous methods that required lengthy…
-
New AI pipeline generates lifelike 3D avatars from single images
Researchers have developed FiCA, a novel system for generating lifelike 3D avatars from a single portrait image. This pipeline combines human-centric vision foundation models with a diffusion model to reconstruct comple…
-
New research enhances 3D Gaussian Splatting for efficiency and editing
Researchers are advancing 3D Gaussian Splatting (3DGS) techniques to improve efficiency, accuracy, and editing capabilities. New methods focus on incorporating uncertainty quantification for better active view selection…
-
New GeoGS-CE method improves channel estimation for high-mobility scenarios
Researchers have developed GeoGS-CE, a novel two-stage framework for channel estimation in high-mobility scenarios like high-speed railways. This method leverages a scene-level 3D Gaussian representation and a different…