ETH3D
PulseAugur coverage of ETH3D — every cluster mentioning ETH3D across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
New EMCStereo method enhances depth estimation for thin structures
Researchers have developed EMCStereo, a novel stereo matching method designed to improve depth estimation for thin structures like tree branches. The method integrates three lightweight attention modules—Efficient Multi…
-
Marigold V2 advances monocular depth estimation using diffusion transformers
Researchers have developed Marigold V2, an advancement in monocular depth estimation that repurposes diffusion transformer (DiT) architectures. This new method achieves sharper and more detailed depth maps by employing …
-
New Gekko method enhances 3D vision pre-training without 3D labels
Researchers have developed a new self-supervised pre-training method called Gekko for 3D computer vision tasks. Gekko leverages the difference in reconstruction error between cross-view completion and masked autoencodin…
-
New method simplifies absolute pose estimation using gravity prior
Researchers have developed a new method for estimating the absolute pose of objects, crucial for robotic applications. This approach leverages gravity direction as prior information, simplifying the 6-DoF problem into a…
-
PhasorNet uses frequency domain for real-time stereo matching
Researchers have developed PhasorNet, a new framework for real-time stereo matching that leverages frequency-domain cues to improve accuracy in challenging scenarios. The system incorporates a Phase-Augmented Transforme…
-
New Self-Geometry pipeline enhances 3D vision model geometric consistency
Researchers have developed Self-Geometry, a novel test-time adaptation pipeline designed to enhance the geometric consistency of 3D vision foundation models. This method directly imposes explicit multi-view geometric co…
-
New protocol assesses 3D reconstruction reliability without ground truth
Researchers have developed a new protocol called "Track-Leakage-Free Hold-Out Self-Validation" to assess the reliability of 3D reconstructions from photogrammetric inspection without external ground truth data. This pro…
-
STEREOFLOW advances stereo matching with generative framework and diffusion transformer · 2 sources tracked
Researchers have introduced STEREOFLOW, a novel generative framework for stereo matching that addresses limitations in traditional deterministic regression approaches. This new method integrates deterministic matching w…
-
New method enhances 3D scene generation using latent space flow matching
Researchers have developed a new method called Latent Riemannian Flow Matching to improve 3D scene generation using geometric foundation models. This technique operates within the latent space of models like the Visual …
-
New EpiDistill method enhances monocular depth estimation using geometric priors
Researchers have developed a new framework called Epipolar Distillation (EpiDistill) to improve monocular depth estimation in AI models. This method transfers scale-aware geometric priors from multi-view models to singl…
-
New stereo matching methods improve accuracy and efficiency · 2-paper roundup
Two new research papers introduce novel approaches to stereo matching, a computer vision task focused on reconstructing 3D scenes from two-dimensional images. WAVE-Stereo proposes a method that combines correlation volu…
-
New SfM methods combine foundation models and depth priors for improved 3D reconstruction · 4 sources tracked
Two new research papers introduce advanced methods for Structure-from-Motion (SfM) reconstruction. Glob3R leverages 3D foundation models and optimizes feed-forward geometric predictions for robust and accurate scene rec…
-
New URS-Stereo framework enhances real-time stereo matching with uncertainty guidance
Researchers have developed URS-Stereo, a novel real-time stereo matching framework designed for applications requiring both speed and accuracy, such as robotics and autonomous systems. The system introduces an Uncertain…
-
New attention mechanisms boost stereo matching accuracy and efficiency
Two new research papers introduce novel attention mechanisms for stereo matching, a computer vision task crucial for 3D reconstruction. The first paper, MatchAttention, embeds explicit matching constraints into attentio…
-
Monocular Depth Estimation via Neural Network with Learnable Algebraic Group and Ring Structures
Researchers have developed LAGRNet, a new framework for monocular depth estimation that incorporates algebraic geometry principles. Unlike previous methods that treat depth estimation as a generic regression problem, LA…