Loading...
Loading...
3D models, rendered datasets, physics simulation, digital twins, synthetic data generation, game engine data
1,029 datasets
600 sequential point clouds acquired at three-hour intervals for monitoring a forested landslide. The data was used to evaluate a temporal aggregation method for improving tree trunk tracking and quantifying 3D landslide displacements. The dataset is associated with a study authored by Ronald Tabernig and was last updated in October 2025.
A 3D global synthetic dataset tailored for seismic wavefield modeling and full-waveform tomography, created by lishiqianhugh and last updated on 2025-09-28. The dataset is organized into acoustic, elastic, and earth model sections, with file sizes ranging from 722 kB to 35 GB per file. It includes multiple HDF5 files containing seismic and wavefield data.
SPAR-7M is a dataset of 7 million synthetic annotations for 3D scenes, created by jasonzhango and hosted on Hugging Face. The dataset was last updated on September 28, 2025. It is derived from third-party datasets including ScanNet and ScanNet++, which must be obtained separately under their respective non-commercial research terms.
Unreal Engine documentation resources up to version 5.3, extracted from HTML files. The dataset includes 409,188 source files and provides original URLs for reference. AdamCodd created this dataset, last updated on June 17, 2024.
SCIOPS collected juvenile salmon, pelagic fish, zooplankton, and oceanographic data from Tucker trawls, CTD, and ADCP units in the coastal Gulf of Alaska. Sampling occurred from July 2001 to November 2004 across four research cruises. The survey covered transects from nearshore to beyond the 200m continental shelf break.
2001-2002 zooplankton species abundance and biomass data collected from vertical net tows in the Gulf of Alaska. The data were gathered by the Gulf of Alaska Long-Term Observation Program using a WP-2 net with a 253-μm mesh, deployed from the F/V Great Pacific. Sampling covered coastal waters and Shelikof Strait, with transects extending from nearshore to beyond the 200-meter shelf break.
RealBokeh 3MP is a dataset presented in the paper 'Bokehlicious: Photorealistic Bokeh Rendering with Controllable Apertures' at ICCV 2025. The dataset likely contains images for training and evaluating photorealistic bokeh rendering models. It was created by Tim Seizinger and collaborators and was last updated on October 16, 2025.
HELVIPAD provides 39,553 paired equirectangular images for omnidirectional stereo depth estimation. The dataset was created by author chcorbi and last updated in September 2025. It includes corresponding pixel-wise depth and disparity labels derived from LiDAR point clouds across diverse scenes.
Snow depth maps with a spatial resolution of 2 meters were generated by digital photogrammetry from ADS 80/100 imagery. The maps represent the difference between a winter Digital Surface Model (DSM) and a summer DSM from 2013, with buildings, vegetation, and outliers masked. The work was published by Bühler et al. in 2015 and last updated in 2019.
A 2020 research dataset from the WAIS Divide ice core site in West Antarctica, produced by AMD_USAPDC. It integrates ice core analyses, borehole deformation measurements, and Electron Backscatter Diffraction (EBSD) microstructure studies. The data aims to link ice fabric evolution, impurity content, and flow dynamics to climate history.
50 Sirius diamict samples were analyzed using micromorphology and bulk texture techniques to determine the character of previous glacial overriding in the Allan Hills nunatak. The dataset includes samples from 40 different sites, all triangulated, photographed, and described, with comparative tillite samples from Mt Feather. The data was collected by SCIOPS and last updated in December 2001.
500,000 synthetic task-oriented dialogues in Traditional Chinese, tailored for Taiwan's social context. The dataset was created by author lianghsun, combining reference-based and reference-free generation methods using LLMs, with a subset derived from the lianghsun/tw-instruct collection. It was last updated on January 10, 2025.
Microsoft's Orca Agentinstruct 1M V1 is a fully synthetic dataset containing approximately 1 million instruction-response pairs. The data was generated using the AgentInstruct framework, which processes raw text content from the public web. The dataset was published by Microsoft in November 2024.
A next-generation dataset for fine-grained and hierarchical 3D part understanding, released for the NeurIPS 2025 Dataset and Benchmark Track. It was created by researchers from ShanghaiTech University and is hosted on Hugging Face. The dataset page was last updated on October 26, 2025.
Synthetic urban scenes generated with SceneCity assets for outdoor inverse rendering and reconstruction. The dataset features diverse urban environments rendered under complex, multi-illumination outdoor lighting conditions to simulate real-world variability.
Key Performance Indicators (KPIs) recorded from gNBs in a simulated Open Radio Access Network (RAN) architecture. The dataset was generated for a digital twin experiment featuring three 5G traffic slices (eMBB, mMTC, URLLC) with both normal and anomalous attack traffic. It was created by Joshua Groen et al. and used for a paper presented at the EuCNC & 6G Summit in March 2024.
RGE ALTI 2015 is a gridded numerical terrain model describing the shape and elevation of the French terrain. It is produced by the Bureau de Recherches Géologiques et Minières and was last updated in March 2019. The model is designed for use at scales between 1:3,000 and 1:20,000 and is available in 1-meter and 5-meter grid resolutions.
A processed version of the 3D-Front dataset, organized into 3D scenes paired with rendered multi-view images and surfaces for the MIDI-3D project. Each scene contains 3D models (.glb), point clouds (.npy), and rendered images in RGB, depth, and normal formats with camera information. The dataset was uploaded by author 'huanngzh' and last updated on 2025-06-25.
5,354 industrial product images for anomaly detection, originally from MVTec and provided by Voxel51. The dataset is formatted for loading with the FiftyOne library, facilitating computer vision tasks in manufacturing inspection.
A multi-camera synchronized video dataset rendered using Unreal Engine 5. It includes synchronized multi-camera videos and their corresponding camera poses. The dataset was released by KlingTeam and updated on April 15, 2025.