Loading...
Loading...
3D models, rendered datasets, physics simulation, digital twins, synthetic data generation, game engine data
1,034 datasets
Kate Mesh provides a coding manual and supplementary examples of gesture forms in still images and video recordings. The materials informed the coding work of the first author and four project reliability coders. The dataset was last updated on March 18, 2024.
New York City's 59 Community Districts contain a 3D model of every building present in 2014, based on a Department of Information Technology and Telecommunications aerial survey. The Department of City Planning enhanced the model by converting it to .3dm format for Rhinoceros 3D and adding base layers like streets and parks. It is compatible with common 3D-modeling software such as SketchUp and AutoCAD.
Structured3D provides a large-scale collection of photo-realistic synthetic indoor scenes for structured 3D modeling, published for ECCV 2020. It contains house designs with ground truth for room layouts and structural annotations.
SensatUrban provides urban-scale photogrammetric point clouds for 3D scene understanding, released by Qingyong Hu in conjunction with CVPR 2021 and IJCV 2022 publications. The data covers large city areas to support benchmarks in semantic segmentation and urban modeling.
ShapeNetVIPC Gen is a dataset of 3D objects, likely containing point cloud data. It was authored by Wang131 and published on the Hugging Face platform, with a last recorded update on 2025-11-17. The dataset's specific scale, columns, and file formats are not detailed in the available metadata.
STL mesh file of a squid, authored by Brady Parrish and hosted by the Texas Data Repository. The record was last updated on March 18, 2024. This 3D model was used in an unspecified analysis.
This collection aggregates state-of-the-art (SOTA) research papers, implementation code, and benchmark datasets for 3D point cloud object detection and semantic segmentation. It organizes deep learning resources to facilitate the development and evaluation of spatial vision models across various 3D computer vision tasks.
Thermal and visible 3D reconstructions acquired using the PyMTI-UAS instrument. The dataset was created by James Thompson during the 2022 Fagradalsfjall volcanic eruption campaign in Iceland and was last updated on March 18, 2024.
Synthetic dataset contains question-and-answer pairs related to PowerShell scripting, Active Directory administration, and Office 365 operations. The dataset was created by author 'adamo1139' and was last updated on the Hugging Face platform in September 2023. Specific details on the number of rows, columns, and data size are not provided.
MVImgNet provides a large-scale collection of real-world multi-view images for 3D vision tasks, published by GAP-LAB-CUHK-SZ at CVPR 2023. The data supports 3D reconstruction, NeRF, and Gaussian Splatting using image sequences of physical objects.
OakInk provides 3D hand-object interaction (HOI) data using the MANO model, released by the OakInk organization in 2022. It serves as a repository for 3D hand pose estimation and motion generation tasks involving human-object manipulation.
Curated by tkuri, this repository provides a bibliographic collection of research papers focused on intrinsic decomposition and inverse rendering as of April 2025. It functions as a specialized index for computer vision researchers seeking peer-reviewed literature and project implementations in scene reconstruction.
70,000 point cloud samples representing handwritten digits across 10 distinct classes. The data is partitioned into 60,000 training and 10,000 test instances derived from 28x28 grayscale images.
A three-dimensional building massing model of New York City, provided by the City of New York's Office of Technology and Innovation. The model is available in three file formats and was last updated in September 2023.
3D point cloud representations stored in HDF5 format featuring exactly 2048 uniformly sampled points per shape. The data provides standardized geometric coordinates for spatial analysis and 3D deep learning applications.
TightCap is a 3D human reconstruction dataset and framework presented at SIGGRAPH 2022 by ChenFengYe. It utilizes multi-view stereo data to capture human shapes under clothing by introducing a specialized clothing tightness field. The data supports the generation of high-fidelity clothed avatars by decoupling body shape from garment geometry.
Presenting a test dataset for GamePhysics, created by author taesiri. The dataset was last updated on January 10, 2024. The specific contents, row count, and column structure are not described.
This dataset supports the 'Three-Filters-to-Normal' method for accurate and fast surface normal estimation, as presented at RAL+ICRA 2021. It is associated with synthetic data, disparity maps, and depth information for training and evaluation. The author is ruirangerfan, and the project was last updated in January 2022.
Developed by the colour-science organization and updated in November 2024, this repository provides spectral data and color space resources for the Mitsuba 3 physically-based renderer. It functions as a specialized collection of datasets for managing color science within spectral rendering workflows.
Tools and data for reconstructing 3D human avatars from unconstrained web-based video data as presented in CVPR 2025. This toolbox facilitates the creation of digital humans from 'in-the-wild' sources rather than controlled studio environments.