Loading...
Loading...
3D models, rendered datasets, physics simulation, digital twins, synthetic data generation, game engine data
1,411 datasets
16,565 furniture meshes constitute the 3D-FUTURE dataset, a collection of 3D models with textures and rendered images. The dataset, hosted by user 'luckfort' on Hugging Face, includes raw OBJ files, materials, textures, and processed subsets for specific workflows. Its last recorded update was on 2026-06-13.
IMOS National Reference Station (NRS) Zooplankton Biomass data comprises monthly biomass sampling from nine stations around the Australian coastline. The plankton ecology lab analyzes samples for community composition, biomass as dry weight, and size spectrum analysis. Data storage is planned to be interoperable through IMOS Infrastructure and is freely available via the Australian Ocean Data Network portal.
A 4.3 MB DOCX file authored by Xiao Ding and last updated on 2026-05-11. The document describes a study investigating the role of integrin α2 in determining the mesenchymal-to-endothelial phenotype transition of human trabecular meshwork cells, relevant to intraocular pressure regulation in glaucoma. Experiments used western blot analysis, immunohistochemistry, collagen contraction assays, and the integrin α2 inhibitor BTT3033.
3DGS-Mesh digital twin environments are released for the HUGE-Bench robotics benchmark. The dataset contains 3D assets required for rendering and collision-aware evaluation, including point clouds and simplified meshes. It was authored by yu781986168 and last updated on 2026-06-24.
739 annotated GUI screenshots are designed for training computer vision models to understand and interact with graphical user interfaces. The dataset uses the specialized COCO4GUI format, which extends the standard COCO detection format to handle GUI-specific features, interaction sequences, and rich metadata. It was created by Voxel51 and last updated on June 11, 2026.
A training dataset for the DiffusionHarmonizer generative AI model developed by NVIDIA. The dataset was curated to support functions like 3D reconstruction artifact removal and object harmonization. The dataset page was last updated on June 24, 2026.
Yiqi Huo's dataset, last updated on 2026-05-14, contains simulation results for topological spin textures in a two-dimensional CrSCl monolayer. The 31.3 MB ZIP file likely includes data from first-principles calculations, Monte Carlo, and atomistic spin simulations. It demonstrates the effects of magnetic field orientation and tensile strain on the stability and switching of skyrmion and bimeron states.
5.5 KB Excel file authored by Wenjun Zhao, last updated in May 2026. This dataset compares the computational costs of exact and entropic approximations for Gromov-Wasserstein and Wasserstein distances when applied to 2D/3D shapes represented as point clouds. It likely contains a structured analysis of algorithmic complexity relative to the number of points per shape (N).
421 stratigraphic section locations across Yukon compiled from 51 published and unpublished surficial geology maps. The dataset includes attribute data for 1198 section layers with unit depths, thicknesses, material classification, and descriptive notes. It was produced by the Geological Survey of Canada and Yukon Geological Survey.
A movie presents multichannel image volumes of three focalcheck beads. The data demonstrates how sonification can reveal subtle spectral differences that are not visually apparent in false color renderings. The dataset was created by Mysore Aprameya and is available under an Open Access license.
Trawl catch data from deepwater commercial trawl nets and fine mesh MIDOC midwater sampling nets deployed from the catcher vessel FV Adriatic Pearl. The dataset contains species, length, weight, sex, and stage information for marine organisms. It is provided by the Australian Ocean Data Network and was last updated on 2026-06-27.
S1 sandstone is a micro-CT image dataset with extracted pore networks, published by Dong and Blunt in 2009. The data is provided by the Imperial College Consortium On Pore-Scale Modelling and includes a header file for visualization in software like Paraview or ImageJ. The dataset is openly accessible under an Open Access (green) license.
Terrestrial LiDAR point cloud data captured by Wollongong City Council surveyors in May 2023 using a Leica P30 scanner. The dataset includes scans of Flagstaff Hill Fort, with underground scans recorded with intensity only due to insufficient lighting. The data has been thinned for presentation purposes.
egnch's LEGO 3D dataset contains 1,418 LEGO sets sourced from the LDraw Official Model Repository. Each set includes 4 rendered views, a colored point cloud with 10,000 points, and 10 English captions. The dataset was last updated on 2026-06-10.
3D models of the brachiopod Rafinesquina capture its shell at different gape angles. The original specimen is reposited with the Cincinnati Museum Center under identifier CMC IP98748. David Peterman of Pennsylvania State University is the author.
A dataset from the Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung supporting the research by Matos et al. (2022). It likely contains calculated oceanographic variables, a spatial mesh, and notebooks for visualization. The dataset is published under an Open Access license.
Aalto University research data for the paper 'Pantomime: Mid-Air Gesture Recognition with Sparse Millimeter-Wave Radar Point Clouds'. The dataset belongs to the Radiosense project and is authored by Sameera Palipana. It is available under an Open Access (green) license.
A set of map images showing the probability of flooding with water depths exceeding 20 cm in the Netherlands. The data was developed by the DPRA and is based on scenarios of primary flood defense breaches from the National Flood Data Database. The dataset is provided by the Dutch Ministry of the Interior and Kingdom Relations under a CC-PDM-1.0 license.
Monitoring data from the Port of Rotterdam Authority tracks the environmental impact and compensation measures for the first phase of Maasvlakte 2, constructed between 2008 and 2013. The dataset likely contains information on measures such as a 25,000-hectare soil protection area and a new dune area to mitigate effects on the Voordelta Natura2000 site. Data was collected under three Monitoring and Evaluation Plans (MEP) to assess the effectiveness of these interventions.
Over 15,000 seabed samples from across the Australian marine region provide sediment properties extracted from Geoscience Australia's Marine Sediment Database. Analytical data includes textural composition, particle size distribution statistics, textural class, and calcium carbonate values where available. The dataset spans coastal, shelf, slope, and deep ocean locations and is maintained by the Australian Ocean Data Network.