Loading...
Loading...
Image-text pairs, instruction tuning, visual QA, cross-modal data, foundation model training data
1,922 datasets
Five consecutive seasons (2020/2021 to 2024/2025) of football data from the five major European leagues, comprising 8,979 matches. The dataset includes detailed match events, player performance metrics, and spatio-temporal shot trajectories. It was created by oualid dehbane and published on figshare under a CC-BY-4.0 license.
The dataset is an evaluation set for multimodal image registration, containing 864 image pairs created from three publicly available 2D datasets. It includes Aerial (Near-Infrared and RGB), Cytological (Fluorescence and Quantitative Phase), and Histological (Second Harmonic Generation and Bright-Field) image pairs, each subjected to four levels of rigid transformations. The data is structured for 3-fold cross-validation, with specific sample counts per fold for each sub-dataset.
206 aligned Bright-Field and Second-Harmonic Generation image pairs, each 834x834 pixels, for evaluating registration methods. The dataset includes a training set of 40 pairs, two validation sets of 25 and 7 pairs, and a test set of 134 pairs with applied random transformations. It was produced by Kevin W. Eliceiri of the University of Wisconsin–Madison, based on data from a breast carcinoma survival study.
A Multimodal Dataset for Mixed Emotion Recognition contains physiological and video data from 73 participants watching emotion-inducing video clips. The dataset includes EEG, GSR, PPG, and frontal face video signals, alongside self-reported ratings on PANAS, VAD, and amusement-disgust dimensions. It was created by Pei Yang from Tsinghua University and shared via Papers with Code.
Experimental data from a study on the visual enhancement of touch effect and its relation to the bodily self, using the rubber hand illusion. The study was conducted by Matthew R. Longo of University College London. Tactile acuity was measured by having participants judge the orientation of square-wave gratings under different visual and stroking conditions.
3,703 ultrasound images from 2,685 patients were used to develop machine learning models for breast cancer diagnosis. The dataset includes 2,069 benign and 616 malignant cases collected between July 2019 and March 2024. Shengxin Pei authored this research, which compares models using BI-RADS terminology, ultrasound imaging, and radiomics features.
A collection of 206 aligned Bright-Field and Second-Harmonic Generation images, each 2048x2048 pixels, for evaluating medical image registration methods. The dataset, produced by Kevin W. Eliceiri of the University of Wisconsin–Madison, is partitioned into training (40 pairs), validation (32 pairs), and test (134 pairs) sets. It originates from research published in Nature Communications Biology on non-disruptive collagen characterization in clinical histopathology.
École Polytechnique Fédérale de Lausanne provides characterization data for research on multimodal host-guest complexation in perovskite photovoltaics. The data includes structural, optoelectronic, and photovoltaic measurements from figures and supplementary information, plus molecular dynamics and DFT calculation files. The dataset supports the findings of the paper with DOI 10.1038/s41467-021-23566-2.
Starting Q1 2017-2018, this table tracks the number of employees delivering multimodal safety and security program oversight for Transport Canada. The data is provided by Statistics Canada and is available in CSV, HTML, and XML formats. It was last updated on 2026-07-06.
Transport Canada's Multimodal Safety and Security Programs Oversight Delivery Indicators contain the number of completed inspections by programs, activities, and administrative regions. The data starts with the four quarters of the 2017-2018 federal government fiscal year. It is published by Statistics Canada on the open_canada platform.
Transport Canada's Multimodal Safety and Security Programs data on immediate risk reduction measures. The table contains the type and number of measures by program, starting from the first quarter of the 2019-2020 federal fiscal year. Statistics Canada is listed as the organization responsible for the data.
Transport Canada's Multimodal Safety and Security Programs Oversight Delivery Indicators include enforcement actions. The data covers the type and number of actions by program, starting from the first quarter of the 2019-2020 federal fiscal year. Statistics Canada publishes this data in CSV, HTML, and XML formats under the OGL-CA-2.0 license.
K-EmoCon is the first publicly available emotion dataset accommodating multiperspective assessment during social interactions. It contains multimodal sensor data from 16 paired debate sessions, each approximately 10 minutes long, with continuous emotion annotations made every 5 seconds. The dataset was created by Cheul Young Park at KAIST and released in 2020.
10 samples constitute this multimodal dataset from Voxel51, hosted on Hugging Face. It is designed for use with the FiftyOne computer vision toolkit. The dataset was last updated on July 16, 2026.
A dataset from 42 individuals who browsed digital supermarket brochures while their neural and ocular activity was recorded. Data includes encephalographic (EEG) recordings, eye tracking (ET) recordings, questionnaire responses, and computer mouse interactions. The dataset was created by Kostas Georgiadis and is available under an Open Access license.
100 healthy control volunteers (80 train, 20 test) provide data for the Big Cross-Modal Attenuation Correction Challenge (BIC-MAC) held in conjunction with MICCAI 2026. The dataset is released by DEPICT-RH and was last updated in July 2026. It includes PET readouts and is associated with a GitHub repository for the project.
GBIF.Org User provides a dataset of 22,292 species occurrence records matching a specific taxonomic query. The data is aggregated from 304 constituent datasets and includes only records with geographic coordinates and no known geospatial issues. Records are limited to specimens or literature occurrences for a defined list of plant taxa, including genera like Caesalpinia and Haematoxylum.
Approximately 500 manually annotated images of Chinese adults, with the dataset last updated on 2026-07-12. The dataset is created by Richarddzz and contains fine-grained textual descriptions and labels for adult content images, intended for research and model training. It is hosted on Hugging Face, with the full dataset requiring contact via Telegram for access.
64 hospital physicians participated in a randomized exploratory trial comparing a 15-minute guided meditative relaxation session to an unguided rest condition. The dataset includes physiological parameters (blood pressure, heart rate) and psychological measures (SPPN, QSCPGS, SRSI3_29) collected at baseline, immediately post-session, and a few hours later. The supplementary file was authored by Siddhiraj Banjac and uploaded to figshare in May 2026.
64 hospital physicians participated in a randomized exploratory trial comparing a 15-minute guided meditative relaxation session to an unguided rest condition. The dataset includes physiological parameters (blood pressure, heart rate) and psychological measures (SPPN, QSCPGS, SRSI3_29) collected at baseline, immediately post-session, and a few hours later. The study was authored by Siddhiraj Banjac and published on figshare in 2026 under a CC-BY-4.0 license.