Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
16,074 datasets
NYC DOT Art collaborates with community-based organizations to commission artists for temporary art installations on city property. The dataset is provided by the City of New York and was last updated on March 8, 2026. Data is available in multiple formats including XML, RDF, JSON, and CSV.
A computer vision dataset named TuSimpleYOLO, hosted on Kaggle. The dataset title suggests a focus on object detection, likely for autonomous driving applications, using the YOLO (You Only Look Once) framework. Specifics regarding size, content, and creation details are unavailable from the provided metadata.
A selection of images and animations explaining the 2004 Indian Ocean tsunami, revised and released by Geoscience Australia for the disaster's tenth anniversary. This collection updates previously released resources and includes content in HTML, PDF, and MP4 formats.
120 magnetic tapes contain water bottle data from 15,000 Russian and foreign oceanographic cruises. The dataset, compiled by the Russian National Oceanographic Data Centre (NODC), includes cruise reference numbers, country, organization, ship, period, and geographical region. This historical collection was last updated in 1985.
Continual-NExT is a benchmark built upon a collection of widely used and publicly available multimodal datasets for both understanding and generation tasks. The benchmark is adopted to evaluate the multimodal continual learning ability of unified generation and understanding MLLMs. It was created by jingyang and last updated on February 19, 2026.
Packed with 94,428 records from the EPA Science Inventory, focusing on reports and presentations. It is one of five archives packaging the same information differently, with records organized by title, entry ID, and page number.
Per-frame face, pose, and hand keypoints extracted from 4,284 Indian Sign Language videos. The dataset likely contains landmark coordinates generated by the Mediapipe Holistic model for each video frame. The author, organization, and specific collection details are unknown.
TeaLeafNet V2 Non Geo Tagged is a labeled image dataset for the early detection of tea leaf diseases. The dataset is hosted on Kaggle, but its author, organization, and specific scale are unknown. Its last update date and licensing information are also unspecified.
JTSGRIT generated this synthetic image dataset in March 2026 using Unity Perception to model pedestrian detection in night fog. It provides fewer than 1,000 images at 1101 x 514 resolution captured from a Kia Ray vehicle's point-of-view at a height of 1.4 meters.
Barnet Libraries records event details, including attendance counts and community outcomes. The dataset lists events facilitated or organized by the library service, with some held outside library premises. It includes columns for event name, date, location, and manual counts of adult and child attendees.
A dataset for training object detection models, likely containing images of vehicles and license plates. The dataset is hosted on Kaggle and is designed for use with the YOLO (You Only Look Once) model architecture. Specific details on the number of images, annotation format, and source are not provided in the available metadata.
A collection of over 46,000 images for classifying car body types into eight distinct categories. The dataset is hosted on Kaggle, but the author, collection method, and specific image sources are unknown. The dataset's age and update history are also unspecified.
Coordinates likely associated with object detection bounding boxes for the YOLOv8 and YOLOv11 model families. The dataset is hosted on Kaggle, but its specific contents, size, and creation details are not provided in the available metadata. Further details such as the number of annotations, the source images, and the author are unknown.
Kaggle hosts a dataset titled 'dataset_resnet_densenet'. The dataset likely contains image data for use with ResNet and DenseNet convolutional neural network architectures. The author, organization, and specific content details are unknown.
Mark Clayton Hand and the Research Cluster on Employee Ownership and Workplace Democracy (EOWD) developed this database of service providers for purpose trusts in the United States. Updated through March 2026, the collection aggregates self-reported data from public sources in partnership with organizations like Project Equity and Purpose Owned.
Anastasiia Vlasenko (Harvard Dataverse) provides this replication dataset and code analyzing the impact of Euromaidan protest activity on decentralization reforms in Ukrainian hromadas. The collection includes data on reform delays, centrally funded grants, and e-democracy participation across local administrative units. It was last updated in March 2026 to support the empirical results reported in the associated research paper.
Comprising parameters from a theoretical redesign of the Fairchild Dornier 728JET aircraft, conducted as part of a 2020 university course. The redesign process used Excel-based tools and NASA OpenVSP for 3D modeling, resulting in calculated deviations from the original aircraft ranging from 0% to 48%. The work includes analysis of lift and drag coefficients, mass predictions, center of gravity, polar curves, and Direct Operating Costs (DOC).
SLAM-YOLO-Pipeline-Scripts is a dataset of scripts likely intended for a computer vision pipeline integrating Simultaneous Localization and Mapping (SLAM) with YOLO-based object detection. It was published on Kaggle, but the specific author, organization, and data content are not detailed. The dataset's size, format, and last update date are unknown.
A collection of images generated by a Projected GAN model, likely using the Deep Feature (DF) variant. The dataset is hosted on Kaggle, but its exact size, creation date, and author are unspecified. The content appears to be synthetic visual data intended for machine learning experimentation.
Kaggle hosts a dataset titled 'projected-gan-scc-generated'. The dataset likely contains synthetic images generated by a Projected Generative Adversarial Network (GAN), possibly for the SCC (Semantic Composition Consistency) task. Its specific content, scale, and creator are not detailed in the provided metadata.