Loading...
Loading...
3D models, rendered datasets, physics simulation, digital twins, synthetic data generation, game engine data
1,034 datasets
HOT3D-Clips is a set of curated sub-sequences from the HOT3D dataset, hosted on Hugging Face by bop-benchmark. The repository was last updated on 2025-09-29. Documentation and Python utilities for loading, undistorting fisheye images, and rendering are provided via the HOT3D Toolkit.
A study conducted with students using the DARA augmented reality tool. Students' emotions were collected via the Emolive facial recognition tool and the AEQ questionnaire. The dataset shows the emotions collected by these two instruments.
NCEP's Limited Fine Mesh model output provides daily tropospheric analyses and 6 to 48-hour forecast fields for North America. The dataset covers a 24-year period from October 1971 through December 1995. Production of these grids was discontinued by NCEP after 1995.
WHU Urban 3D provides point cloud data for urban scenes, likely for 3D reconstruction and modeling. The dataset, uploaded by user 'astroy' to Hugging Face, was last updated on July 31, 2023. A dedicated Python library, pywhu3d, is available for interacting with the data.
The HIT dataset contains 398 paired observations of internal body tissues and external surface shapes, released by varora in 2024. It features volumetric segmented MRI scans identifying bones, lean, and adipose tissue alongside corresponding SMPL meshes for 157 male and 241 female subjects.
Synthetic text data for classification tasks, generated using the Distilabel framework. The dataset includes a reproducible pipeline configuration file (pipeline.yaml). Specific row count, column count, and file size are not detailed in the provided input.
The DTD dataset contains 5,640 texture images organized into 47 categories based on human perception, with 120 images per category. Images were collected from Google and Flickr and annotated via Amazon Mechanical Turk, with each image assigned a main category and joint attributes. Image sizes range from 300x300 to 640x640 pixels.
A geospatial dataset describing the development plan for the 'Blumenstrasse Im Almeshofen' area in the city of Püttlingen, Germany. The data is provided via an INSPIRE-compliant ATOM feed download service, with links generated dynamically from a WMS layer. The dataset was last updated on January 10, 2024, and is managed by the German Federal Agency for Cartography and Geodesy (BKG).
Over 1,000 color image textures at 512x512 resolution with associated English text descriptions. The dataset was created by the author 'dream-textures' for training and fine-tuning diffusion models on texture generation tasks. It was last updated on January 13, 2023.
348,566 MEDLINE references from 270 medical journals published between 1987 and 1991. Each record contains titles, abstracts, and Medical Subject Headings (MeSH) indexing terms alongside author and publication type metadata.
A 3D model of an astronaut created by the National Aeronautics and Space Administration, containing 6,381 polygons and 3,739 vertices. The model is available in the IMAGE/X-3DS file format and was last updated in April 2025.
A collection of texture images intended for training text-to-image or other generative models. The dataset is provided by JapanDegitalMaterial and was last updated on 2023-09-23. All materials used in the dataset are stated to be under a CC0 (Public Domain) license, aiming to avoid copyright issues.
155 surficial sediment samples from Baltimore Harbor's Patapsco and Back Rivers provide data on physical properties, particularly grain size. The Maryland Geological Survey collected the samples during three field campaigns in November 1994, February 1995, and June 1996. This dataset offers a snapshot of sediment characteristics in an urban estuary.
ASO L4 Lidar Point Cloud Digital Terrain Model 3m UTM Grid V001 provides 3-meter resolution gridded bare-earth elevation data, excluding trees. The dataset was created by the NASA/JPL Airborne Snow Observatory aircraft survey campaigns and last updated in October 2019.
ShapeNet-MPC is a dataset for point cloud completion proposed by the research paper 'DuInNet: Dual-Modality Feature Interaction for Point Cloud Completion'. The dataset was uploaded by author xinpuliu to Hugging Face and was last updated on March 1, 2025. It contains 3D object data organized by category, with files for both color images and point clouds.
ModelNet-MPC is a dataset for point cloud completion proposed by the paper 'DuInNet: Dual-Modality Feature Interaction for Point Cloud Completion'. The dataset, authored by xinpuliu, was last updated on the Hugging Face platform on March 1, 2025. It contains directories for ground truth (GT), images (Img), and partial point clouds with and without noise.
The ScanNet dataset provides richly-annotated 3D reconstructions of indoor scenes for computer vision tasks. It is designed for 3D object detection and 3D segmentation applications. The dataset was created by researchers including Angela Dai and Angel X. Chang, with foundational work published at CVPR in 2017.
Texture data from OpenML provides a resource for analyzing visual patterns and surface properties. The dataset's author, organization, and specific size are not documented. It is available through the OpenML platform.
Soil texture data extracted from a multi-layer soil characteristics database for the conterminous United States. The data were generated for regional study areas and are representative of conditions during the Soil Moisture Active Passive Validation Experiment 2008 (SMAPVEX08) campaign. The dataset is hosted by NASA Earthdata and was last updated by NSIDC_CPRD in October 2008.
Soil texture classification data derived from field surveys conducted during the Soil Moisture Active Passive Validation Experiment 2012 (SMAPVEX12). The map provides information about vegetation present in the study area. The data was produced by NSIDC_CPRD and last updated on July 19, 2012.