Loading...
Loading...
3D models, rendered datasets, physics simulation, digital twins, synthetic data generation, game engine data
1,402 datasets
Electron backscatter diffraction (EBSD) measurements of Surfalex HF (AA6016A), an aluminium alloy used in automotive applications. The data was collected by Sumeet Mishra at the University of Manchester using an FEI Sirion field emission gun scanning electron microscope with a Nordlys detector. Scans were performed on three different planes (RD-TD, RD-ND, TD-ND) with an acceleration voltage of 20 kV, a working distance of 16 mm, and a step size of 2 μm.
A pistachio orchard in Spain was surveyed on 29 July 2021. The dataset comprises 248 RGB images captured from a UAV at two different gimbal angles, alongside derived products including ground control points, a 3D point cloud, DEM, and orthomosaic. It was created by Sergio Vélez of Wageningen University & Research for photogrammetric reconstruction in precision agriculture.
LiDAR point clouds from various airborne acquisition projects conducted across Canada. The data is part of the CanElevation Series created by Natural Resources Canada to support the National Elevation Data Strategy. Point cloud files are distributed in the compressed .LAZ / Cloud Optimized Point Cloud (COPC) format and are referenced to the NAD83(CSRS) and CGVD2013 datums.
Geoscience Australia maintains a national collection of marine geological samples and analytical data. This standalone dataset collates sediment properties for over 15,000 seabed samples from the Marine Sediment Database (MARS), which holds approximately 2.6 million digital records. Analytical data includes textural composition, particle size statistics, and calcium carbonate values, with sample depth, location, and survey information.
The Paterson National Geoscience Agreement project used 3D Geomodeller software to construct a geologically-constrained three-dimensional volumetric model of the northwest Paterson Orogen in Western Australia. The model focuses on the Cottesloe Syncline district and was built by project members and specialists from Geoscience Australia. The resultant model, including sections, maps, and images, was exported to Virtual Reality Modelling Language (VRML) format for broad accessibility.
Six topographic models of a natural stone step-pool unit after FAVORization in FLOW3D, each corresponding to a specific flow rate. The models are stored as .stl files with a mesh resolution of 2.5 mm in X, Y, and Z directions. An accompanying Excel file contains inlet boundary locations, water levels, and flow velocities for all numerical models, authored by Chendi Zhang of Tsinghua University.
A dataset supporting research on texture-enhanced thermography for joint inflammation detection. It contains 2,754 thermal images in CSV format across 14 anatomical regions, along with patient metadata and extracted statistical features. The data was collected by researchers including Dima Bykhovsky and published in 2026.
Five datasets of FIB-SEM tomograms showing the steel-concrete interface in mortar and concrete specimens. The datasets include NCI-1, NCI-2, NCI-3, NCI-4, and CI, with voxel sizes of 30 nm or 50 nm. They were acquired by Nicolas Ruffray at ETH Zurich.
Six distinct chemical lists compiled by researchers at the University of Luxembourg support suspect screening for neurotoxicants. The lists range from 33 to 1,243 entries and include documented neurotoxicants, potential negative controls, and literature-mined candidates. Emma Schymanski led the work, which is described in a paper currently under revision.
IHTAclassroom provides visual and acoustic 3D architectural models of a lecture classroom at the Institute of Hearing Technology and Acoustics (IHTA), RWTH Aachen University. The dataset includes geometry, materials, and light sources at multiple levels of detail (LOD) for visual cues, and high-LOD geometry for acoustic cues. It was created by Josep Llorca-Bofí and used in a 2021 publication in Frontiers in Built Environment.
IHTApark is a multi-detailed 3D architectural model of a park, specifically designed for sound perception research in virtual reality. It integrates visual photogrammetry and acoustic models, with data available separately for visual and acoustic cues at a high level of detail. The dataset includes geometry in FBX format, visual textures in JPEG, and a pre-configured virtual reality scene in Unreal Engine.
Gesture recognition and animation data from the CHI 2020 paper "Touché: Data-Driven Interactive Sword Fighting in Virtual Reality". The repository includes separate archives for gesture recognition data and code, animation data and code, and user study results. Data from user studies is provided in JASP files for statistical analysis.
Aalto University researcher Thomas McKenzie measured spatial room impulse responses in a variable acoustics room. The dataset uses two spherical microphone arrays across three source and seven receiver positions, repeated under five different room acoustic configurations. Possible applications include six degrees-of-freedom analysis and rendering, SRIR interpolation, and spatial dereverberation.
Mancos shale samples from The University of Texas at Austin were scanned using X-ray tomography before and after triaxial testing. Testing involved various stress loading paths and confining pressures from 100 to 500 psi on three sample types with different bedding orientations. The dataset includes videos of fracture network visualizations and images scanned at 4 fps with 2984 projections over 360 degrees, with resolution ranging from 65 to 75 microns.
Four supplemental data sets describe deltaic sand deposits at the head of Lewis and Clark Lake, Nebraska-South Dakota, investigated for suitability as a proppant feedstock. Data includes core lithology descriptions in CSV, apparent resistivity profiles in PNG, laboratory reports on particle size and shape in PDF, and sample photomicrographs in JPEG. The dataset was authored by Brent M. Hall of the United States Geological Survey and relates to a 2015 investigation.
A dataset containing quantitative evaluation results for real-scene 3D models simplified at rates of 0%, 20%, 40%, 60%, 80%, and 90%. The data includes image-based similarity metrics, Hausdorff distance results, positional accuracy measurements, and HCI-based evaluation scores from six evaluators. It was authored by HONG WEI and last updated on June 1, 2026.
Curie point depth mapping across Yukon, Canada, was produced using public aeromagnetic data from Natural Resources Canada. The dataset, published by the Government of Yukon, demarcates 95% of the territory into regions of shallow or deep Curie point depth, indicating higher or lower crustal heat flow. Results from two different methodologies and window sizes were broadly consistent, identifying south-central Yukon as having higher heat flow.
5.5 KB of experimental results from an ablation study of a Digital Twin component for urban traffic networks. The dataset, created by Fayez Alanazi and last updated in June 2026, supports a framework integrating Graph Convolutional Networks and Long Short-Term Memory units for traffic forecasting with a reinforcement learning module for adaptive signal control.
Geoscience Australia and the Paterson National Geoscience Agreement project constructed a three-dimensional volumetric model of the Cottesloe Syncline district in the northwest Paterson Orogen, Western Australia. The model was built using 3D Geomodeller software to test geological hypotheses and was exported in Virtual Reality Modelling Language (VRML) format for broad accessibility. The associated report documents the model building capability and geological insights gained from the exercise.
8,544 macroscopic images of wood species form this dataset, created by researchers from Imperial College London and Aristotle University of Thessaloniki. The dataset supports wood species recognition through multidimensional texture analysis, as described in a 2018 research paper. It is version 2 of the dataset, with the underlying data archived on Zenodo in 2019.