Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
15,432 datasets
Polycyclic aromatic hydrocarbon distributions in McArthur River ore samples indicate hydrothermal fluid temperatures of 250-400°C, revising previous lower estimates. The data provides evidence for ore formation in a marine diagenetic environment and serves as a repository for paleobiological information. Findings are based on research submitted by Chen et al. and published by Logan et al. in 2001.
A video file from the Biomedical Research Imaging Center at UNC-Chapel Hill, acquired on a 9.4T Bruker MRI scanner. The dataset shows a dynamic PETALUTE imaging sequence capturing approximately 10 seconds of signal propagation following a hyperpolarized [1-¹³C]pyruvate injection into a phantom. The data was uploaded by author Megan Pike in April 2026.
Naoto Ienaga published a dataset on figshare in April 2026 detailing the percentage of low-confidence keypoints in pose estimation. The data, stored in a 5.5 KB XLS file, compares confidence values below 0.3 for keypoints detected by RTM-tri and MP-tri models, with shoulder keypoints present in all frames.
ParaVT provides source media archives for the ParaVT training corpus. The dataset bundles raw video files referenced by the ParaVT-Parquet annotations, packaged as per-source zip archives. It was created by ParaVT and last updated on May 18, 2026.
ParaVT-Parquet provides two training corpora for a multi-agent framework designed for long-video understanding. The dataset includes a 97,000-row corpus for supervised fine-tuning and a 4,406-row corpus for reinforcement learning. The ParaVT project created this data, which was last updated on May 18, 2026.
Ukrainian administrative data on criminal offenses committed at enterprises, institutions, and organizations, categorized by type of economic activity. The data includes results of pre-trial investigations and information on persons who committed the offenses. It was last updated on 2026-05-07 and originates from the Prosecutor General's Office of Ukraine.
Annotations for 3D assets used in LychSim, a controllable and interactive simulation framework for vision research. The dataset provides semantic category, canonical scale, pose alignment, and mesh offset for each asset. It was created by author wufeim and was last updated on May 14, 2026.
530.7 MB of supplementary code and genome assembly features supporting the 2026 manuscript by Despot-Slade et al. The dataset, hosted on figshare, is intended to reproduce results on the relationship between reproductive mode, ploidy, and centromere organization in plant-parasitic nematodes. It was last updated on May 4, 2026, under a CC-BY-4.0 license.
Supplementary Material 7 from a study on immuno-competent bladder cancer organoids. The 36.8 KB XLSX file, published by Carlotta Frascolla on figshare under a CC-BY-4.0 license, likely contains supporting data for ex vivo drug screening experiments. Its specific content and row count are not detailed in the available metadata.
Aéro-Photo captured a high-resolution aerial photograph of the Lac-Mégantic train derailment site on July 10, 2013. The image is georeferenced and provided as a WMS web service to identify the accident's impact radius. Raw image access requires contacting Aéro-Photo (1961) inc.
U.S. survey experiment data collected via Prolific's nationally representative sample and the Qualtrics platform. The dataset supports research on war propaganda, normative framing, and side-taking. It was authored by Hanson, Kolby and last updated on June 5, —.
A 2018 UNESCO/IOC workshop report synthesizes expert-identified worst-case seismic tsunami scenarios for the Tonga-Kermadec trench. Fourteen experts from four countries ranked scenarios by probability and discussed gaps in instrumentation and community readiness. The report, hosted by the Australian Ocean Data Network, presents these findings and recommendations.
UCDP's most disaggregated dataset covers individual events of organized violence in South Africa. Events are geo-coded down to the level of individual villages and have temporal durations disaggregated to single days. The dataset is based on research from Uppsala University and was last updated on 2026-04 03.
UCDP's most disaggregated dataset covers individual events of organized violence in India. Events are geo-coded to the level of individual villages and have temporal durations disaggregated to single days. The dataset is based on research from Uppsala University and is licensed under CC-BY-3.0-IGO.
Aalborg University researcher Thomas Arildsen presents a critical perspective on the academic social network ResearchGate. The author details personal reservations about the platform while explaining a strategy of internal critique. The dataset likely contains textual analysis or commentary on academic networking practices.
A dataset from the Australian Ocean Data Network presents fossil analysis from a 47 cm sediment core documenting seabed community succession beneath the Amery Ice Shelf. The data reveals a transition from a barren state prior to ~9600 years ago to a diverse modern fauna dominated by filter feeders and polychaetes, linked to ice shelf retreat and advected food supply. This record, last updated in 2026, provides evidence for understanding how sub-ice shelf ecosystems respond to climate-driven changes.
Eliška Rezlerová from the Czech Academy of Sciences authored a computational study providing molecular-level insight into shale gas adsorption and transport. The manuscript describes atomistic simulations generating realistic models of mature and overmature type-II kerogens with micro- and mesoporosity. These models include components like asphaltene, resin, hydrocarbons, and carbon dioxide/water fractions, which are typically not modeled, and simulate the adsorption and diffusion behavior of methane and methane/carbon dioxide mixtures.
A geospatial dataset from the Government Digital Service identifying zones within the town of Nantwich, Cheshire, where well-preserved, waterlogged archaeological deposits are known to survive. The description notes the deposits are of national importance, with preservation quality comparable to sites in Carlisle, York, and London. A management strategy for these deposits is referenced on the Cheshire Archaeology website.
Geoscience Australia revised and released a selection of explanatory animations and illustrations on the tenth anniversary of the 2004 Indian Ocean tsunami. The resources update existing materials and were issued for media and interested organizations. The collection includes PDF, MP4, and HTML files.
A hydrogeological inventory for the Canning Basin in Australia, aggregated by the Australian Ocean Data Network and last updated on 2026-04-10. The dataset contains descriptive attributes grouped into themes including location, geology, hydrogeology, groundwater management, and land use. It describes a basin with sedimentary rocks over 15,000 m thick, a complex depositional history from the Early Ordovician to Early Cretaceous, and includes information from around 250 petroleum wells.