Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
16,527 datasets
Intel Image Dataset is a collection of images for classifying scenes as natural or urban. The dataset is hosted on Kaggle and is associated with computer vision tasks. Specific details on size, collection method, and timeliness are not provided in the available metadata.
A benchmark dataset for multispectral object detection. The dataset is published on Kaggle and likely contains paired image data from multiple spectral bands. The author, organization, and specific scale are unknown.
A collection of medical images for skin disease classification. The dataset is published on Kaggle and likely contains images organized into 10 distinct diagnostic categories. Details on the collection source, time period, and specific conditions are not provided in the available metadata.
SteadyDancer is a dataset for human image animation, focusing on harmonized and coherent motion generation while preserving the first frame. It was created by researchers from Nanjing University and Tencent and was last updated on November 27, 2025. The dataset likely contains sequences of human images or videos for training and evaluating animation models.
Orthographic aerial photography tiles provide georeferenced imagery for the city of Winnipeg. The dataset includes links to image files with associated metadata such as Year, Season, Resolution, and Location. It is published by data.winnipeg.ca and was last updated in October 2025.
Niantic Labs published this benchmark in 2022 for the ECCV paper 'Map-free Visual Relocalization,' which focuses on estimating metric pose relative to a single reference image. The dataset provides a standardized evaluation framework for visual relocalization tasks that do not rely on pre-existing 3D maps or scene-specific training. It includes image pairs and ground truth 6DoF relative poses to facilitate research in map-free navigation.
Annual cost report data from home health agencies, including facility characteristics, utilization, and financial statements. The dataset is organized by CMS Certification Number and was last updated in December 2025. It is provided by the U.S. Department of Health & Human Services.
A public use file provides select measures from skilled nursing facility annual cost reports. It includes provider information, facility characteristics, utilization data, cost and charges by cost center, Medicare settlement data, and financial statement data organized by CMS Certification Number. The dataset is produced by the U.S. Department of Health & Human Services and was last updated in December 2025.
StreamingQA-120K is a large-scale dataset containing 1.28 million videos sourced from WebVid, InternVid, and Panda-70M. It was created by the author 'interlive' and last updated in December 2025. Videos are grouped by similarity for evaluating models on streaming video understanding tasks.
4,272 individual email messages extracted from screenshot images using the Qwen 2.5 VL 72B vision model. The dataset was created by 'to-be' and last updated on December 4, 2025. The source images are from the U.S. House Oversight Committee release.
323,060 anime images are included in this webdataset, resized to a maximum dimension of 640 pixels. The dataset is split into training, validation, and test sets, with the training set containing 257,996 images. It was created by 'just-a-try' and last updated on Hugging Face in November 2025.
City of New York E-ZPass Readers 2021-July 2024 dataset contains traffic speed data collected from readers installed throughout NYC streets. The data, ranging from April 8, 2021 to July 7, 2024, was used to plan traffic calming measures and reorganizations, including projects like the Midtown in Motion congestion management program.
25,197 manually labeled images form a scene text detection dataset. The collection contains mostly Latin and Russian text. It was created by author DonkeySmall and was last updated in December 2025.
TopAir provides video frames captured from a nadir view using the AirSim simulator in various outdoor environments. The dataset includes corresponding depth maps, semantic segmentation maps, and extrinsic camera locations for each frame. It was curated by Yara AlaaEldin and shared by MaLga Lab, with a last recorded update on 2025-11 18.
1.9 million images from the Open Images V7 project annotated with bounding boxes and captions. The annotations were generated automatically using image-text-to-text models. The dataset was uploaded by Fhrozen to Hugging Face and last updated on November 17, 2025.
X-Ego-CS is a multi-agent gameplay video dataset for cross-egocentric multi-agent video understanding in Counter-Strike:2. It contains 124 hours of synchronized first-person gameplay footage captured from 45 players. The dataset was authored by wangyz1999 and last updated on 2025-11-12.
7 subsets of image-text pairs feature exhaustive mask annotations for every object instance corresponding to specific Noun Phrases (NPs). These image annotations are verified by a consensus of three human reviewers to provide a high-precision benchmark for promptable concept segmentation (PCS).
María Sanz's dataset from e-cienciaDatos Harvested Dataverse, last updated in November 2025, presents experimental results for producing jet-fuel precursors. The data covers the aldol condensation of furfural with methyl-isobutyl ketone over defect-engineered UiO-66(Zr) catalysts. Catalyst characterization includes XRD, argon adsorption, TGA, TEM, and FTIR spectroscopy results.
Infinity-Doc-400K contains 400,000 real-world and synthetic scanned documents featuring diverse layout variations and structural annotations, released by infly in 2025. This dataset expands upon the previous 55K version to support document parsing model training across financial, medical, and academic domains. It is provided in Parquet format and includes both image and text modalities.
Metadata for the VGGSound dataset, annotated with video and audio prompts for the paper "Does Hearing Help Seeing? Investigating Audio-Video Joint Denoising for Video Generation". The dataset was created by author jianzongwu and the metadata was last updated on December 3, 2025. The repository contains annotated train and evaluation metadata, while the video files must be downloaded from a separate source.