Depth Anything 3
PulseAugur coverage of Depth Anything 3 — every cluster mentioning Depth Anything 3 across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
New diffusion models tackle in-the-wild video shadow removal
Researchers have developed two new methods for removing shadows from videos using diffusion models. WildShadowRemover adapts a pre-trained video diffusion model with a detail injection module and a frequency-decomposed …
-
AI reconstructs 3D surgical scenes from microscope images
Researchers have developed a method to reconstruct millimeter-range 3D surfaces of neurosurgical operative exposures using standard monocular operating-microscope images and microscope pose data. The technique leverages…
-
RayTun3R adapts 3D foundation models for fisheye cameras
Researchers have developed RayTun3R, a novel method to adapt existing 3D foundation models for use with fisheye camera imagery. These models, which typically perform well with standard pinhole cameras, degrade significa…
-
S-Agent framework enhances VLMs for 3D spatial reasoning · 4 sources tracked
Researchers have introduced S-Agent, a novel framework designed to enhance visual language models (VLMs) for spatial reasoning in 3D environments. S-Agent integrates temporal memory and a hierarchy of spatial tools to e…
-
PaGeR framework adapts 3D models for 360-degree panoramic scene reconstruction
Researchers have developed PaGeR, a framework that adapts existing 3D foundation models, originally designed for perspective images, to reconstruct full 360-degree scenes from single panoramic images. This approach allo…
-
Spark3R accelerates 3D reconstruction with asymmetric token reduction
Researchers have developed Spark3R, a novel framework designed to accelerate feed-forward 3D reconstruction models that utilize Vision Transformers. The method addresses the computational challenge posed by processing e…
-
Depth prior enhances robot navigation through glass surfaces
Researchers have developed a new framework to improve robot navigation in environments with glass surfaces. This method utilizes depth foundation models as a structural prior, aligning them with raw sensor depth data us…
-
AirZoo dataset offers large-scale aerial 3D vision training data
Researchers have introduced AirZoo, a large-scale dataset designed to address the scarcity of training data for aerial geometric 3D vision tasks. The dataset features a scalable generation pipeline using 3D meshes, exte…