Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
15,906 datasets
Antarctica's Vestfold Hills region provides 41 water samples from 10 saline lakes. The dataset contains measured and calculated parameters for major ions, density, and salinity. It was published by the Australian Antarctic Data Centre in 1989.
Geochemical data from eight major shale formations in Alberta provides measurements for Rock Eval pyrolysis and total organic carbon to assess hydrocarbon potential. The Government of Alberta compiled this tabular dataset, which was last updated in March 2026. It supports analysis of shale gas, oil, and liquids resources.
April 1992 to January 1993 current meter data includes velocity and temperature readings from various depths. The dataset was collected by SAIC for the California Monitoring Program (CAMP), funded by the Minerals Management Service, near oil platform Hidalgo. It is hosted by the Center for Coastal Studies at Scripps Institution of Oceanography.
13,000+ eye images labeled for cataract detection, organized into three classes. The dataset was designed for mobile screening applications and includes benchmark results from five convolutional neural networks. It is hosted on Kaggle.
AlcoVision is a dataset for training computer vision models to detect alcohol-related objects. The description indicates it is built for high-precision detection using the YOLOv8 architecture. Specific details on data volume, source, and creation date are not provided in the input.
NASA's Solar Dynamics Observatory provides 4-Hz integrations of solar extreme ultraviolet irradiance data, beginning in 2010 and continuing. The Level 1 Version 8 data products are created by the Laboratory for Atmospheric and Space Physics and include updated long-term degradation corrections. This release from the Science Processing and Operations Center replaces all previous versions.
Funds_handwritten_dataset is a collection of images for optical character recognition tasks, published on Kaggle. The raw description indicates it is intended for OCR tests and model fine-tuning. Specific details on the number of images, their source, and creation date are not provided in the available metadata.
VARISHTA-MM50 appears to be a medical imaging collection, likely containing 50 samples or scans. It is organized and hosted on Kaggle, a platform for data science projects. The specific medical modality and annotation details are not provided in the available metadata.
A dataset from UNICEF Data and Analytics (HQ) measuring national compliance with International Labour Organization standards for paid maternity leave. It indicates whether a country's law provides for 14 weeks or more of paid leave. The dataset was last updated on 2026-03-26.
A public opinion poll conducted in Michigan by Oliver Quayle. The dataset was last updated on 2026-05-21 and is hosted by the Roper Harvested Dataverse. Its specific temporal coverage and content details are not provided in the available metadata.
Fossil specimens from a 1953 fire were documented using singed photographs. This collection substantiates earlier stratigraphic work on marine fossils from the Kimmeridgian, Tithonian, Neocomian, and Aptian stages in Western Australia. The data originates from publications by Brunnschweiler, Guppy, Fairbridge, and others, compiled by Geoscience Australia.
Geoscience Australia presents an environmental analysis of an aeolian deposit within the Jurassic Jurgurra Sandstone, based on a 1976 field examination. The report details a 5-meter-thick exposure near Geegully Creek, correlating it with the subsurface Wallal Sandstone unit, which reaches 369 meters in thickness.
A research dataset from a study examining the influences of employee diversity climate perceptions and quiet quitting on interpersonal deviance. The dataset likely contains survey responses measuring these psychological and behavioral constructs. It was authored by Talukder, Md Farid and harvested by the Texas Data Repository on 2026-04-27.
Spectrum reports from Scaffold PTM software for mass spectrometry analysis of the plant receptor kinase ERECTA and its co-receptor BAK1. The data supports the manuscript "Preventing Inappropriate Signals Pre- and Post-Ligand Perception by a Toggle-Switch Mechanism of ERECTA". It was authored by Keiko U. Torii and harvested from the Texas Data Repository on Dataverse.
More than 2,500 coastal Douglas-fir trees from 30 populations were sampled across multiple common gardens over three years to quantify foliar fungal loads. The dataset, created by Oregon State University and last updated in 2026, uses a hamPCR metabarcoding technique to measure abundance indices for fungal taxa, bypassing issues of compositional data. It captures interactions between host tree ecotype, garden environment, and annual variation on fungal community assembly.
2000 manganese nodules were recovered from the Indian Ocean floor during a 1976 research cruise. This dataset documents a deposit estimated to cover 900,000 km², discovered via bottom photography and sampled by HMAS Diamantina.
A Data Management and Sharing Plan outlines the scientific data to be generated and/or used in a research project. The plan describes a strategy for managing and sharing project data related to the identification of first-in-class ligands that bind glial fibrillary acidic protein (GFAP). It was authored by Alison Axtman and last updated on May 11, 2026.
A Data Management and Sharing Plan outlines the scientific data to be generated and/or used in a research project and describes a strategy for managing and sharing that data. The plan is authored by Kevin Weeks and originates from the ODUM Harvested Dataverse. It was last updated on May 11, 2026.
A Data Management and Sharing Plan authored by Bethany Hedt-Gauthier, last updated on 2026-05-11. It describes the scientific data to be generated and/or used in the research project 'Sex, HIV, and Lung Health Across the Life Course: The Uganda Lung Health Study'. The plan outlines a strategy for managing and sharing the project's data.
Yearly data from the Electronic Labor Organization Reporting System (e-LORS), established under the Labor-Management Reporting and Disclosure Act. The system facilitates the electronic filing, storage, and disclosure of data submitted to the Department of Labor by labor unions, employers, and other entities. The dataset was last updated on March 7, 2026.