Dataset Card for Open Images Dataset
This dataset contains images from the Open Images dataset. It includes image URLs, split into training, validation, and test sets.
Dataset Details
Dataset Description
Open Images is a dataset of approximately 9 million URLs to images that have been annotated with image-level labels, bounding boxes, object segmentation masks, and visual relationships.
Curated by: Google LLC License: Images: CC BY 2.0 license… See the full description on the dataset page: https://huggingface.co/datasets/bitmind/open-images-v7.
Attribution 4.0 (CC BY 4.0)https://creativecommons.org/licenses/by/4.0/
License information was derived automatically
## Overview
Open Images V7 is a dataset for object detection tasks - it contains Objects annotations for 1,892,276 images.
## Getting Started
You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.
## License
This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).
bitmind/open-images-v7-subset dataset hosted on Hugging Face and contributed by the HF Datasets community
abcd10987/open-images-v7 dataset hosted on Hugging Face and contributed by the HF Datasets community
Attribution-ShareAlike 4.0 (CC BY-SA 4.0)https://creativecommons.org/licenses/by-sa/4.0/
License information was derived automatically
Labeled datasets are useful in machine learning research.
This public dataset contains approximately 9 million URLs and metadata for images that have been annotated with labels spanning more than 6,000 categories.
Tables: 1) annotations_bbox 2) dict 3) images 4) labels
Update Frequency: Quarterly
Fork this kernel to get started.
https://bigquery.cloud.google.com/dataset/bigquery-public-data:open_images
https://cloud.google.com/bigquery/public-data/openimages
APA-style citation: Google Research (2016). The Open Images dataset [Image urls and labels]. Available from github: https://github.com/openimages/dataset.
Use: The annotations are licensed by Google Inc. under CC BY 4.0 license.
The images referenced in the dataset are listed as having a CC BY 2.0 license. Note: while we tried to identify images that are licensed under a Creative Commons Attribution license, we make no representations or warranties regarding the license status of each image and you should verify the license for each image yourself.
Banner Photo by Mattias Diesel from Unsplash.
Which labels are in the dataset? Which labels have "bus" in their display names? How many images of a trolleybus are in the dataset? What are some landing pages of images with a trolleybus? Which images with cherries are in the training set?
Open Images Dataset V7 (test set)
Original paper: A Step Toward More Inclusive People Annotations for Fairness Homepage: https://storage.googleapis.com/openimages/web/extended.html Bibtex: @inproceedings{miap_aies, title = {A Step Toward More Inclusive People Annotations for Fairness}, author = {Candice Schumann and Susanna Ricco and Utsav Prabhu and Vittorio Ferrari and Caroline Rebecca Pantofaru}, booktitle = {Proceedings of the AAAI/ACM Conference on AI, Ethics… See the full description on the dataset page: https://huggingface.co/datasets/nlphuji/open_images_dataset_v7.
bitmind/open-image-v7-256 dataset hosted on Hugging Face and contributed by the HF Datasets community
EiMon724/bitmind-open-image-v7-256 dataset hosted on Hugging Face and contributed by the HF Datasets community
"Wake Vision" is a large, high-quality dataset featuring over 6 million images, significantly exceeding the scale and diversity of current tinyML datasets (100x). The dataset contains images with annotations of whether each image contains a person. Additionally, the dataset incorporates a comprehensive fine-grained benchmark to assess fairness and robustness, covering perceived gender, perceived age, subject distance, lighting conditions, and depictions. This dataset hosted on Harvard Dataverse contains images, CSV files, and code to generate a Wake Vision TensorFlow Dataset. We publish the annotations of this dataset under a CC BY 4.0 license. All images in the dataset are from the Open Images v7 dataset, which are sourced images from Flickr and are listed as having a CC BY 2.0 license.
Attribution 4.0 (CC BY 4.0)https://creativecommons.org/licenses/by/4.0/
License information was derived automatically
Media-Text dataset comprising images of banners, posters, covers and another images characterised for media industry.
Full paper is available here: https://www.researchgate.net/publication/385351709_Media-Text_a_Media_Industry-Based_Dataset_for_Scene_Text_Detection" target="_blank" rel="noopener">Media-Text: a Media Industry-Based Dataset for Scene Text Detection
Annotation Format - Each image has corresponding gt_*.txt file, which contains annotations in bounding box format (defined by 4 courners), transcription, and bool flag which determines that text is illegible for OCR. Proposed format is similar to ICDAR15 annotations.
x1, x2, ..., x4, y4, transcription, OCR Flag
Example:
37,68,198,49,214,181,52,200,LADIES,False
ACKNOWLEDGMENT
This work was supported by the Silesian University of Technology (SUT) through the subsidy for maintaining and developing research potential grant in 2024 for young researchers, No. 2/070/BKM24/0058, and by the Ministry of Science and Higher Education "Implementation Doctorate" No. DWD/5/0511/2021.
Thanks to the graphic department of media-press group for the preparation and possibility of sharing graphics thematically related to the prepared dataset.
LICENSE
Annotations created by authors are licesned under CC-BY-4.0 license.Images from the Open-Image-V7 dataset and are licensed according to their source information. Source information is defined in a file metadata.csv file that defines all the metadata of each file (File name corresponds to the ImageID column).
Images whose name corresponds to the media_press pattern are provided for academic use.
</div>
<div>@inproceedings{inproceedings,</div>
<div>author = {Kalisz, Seweryn and Marczyk, Michał and Polanska, Joanna},</div>
<div>booktitle = {Modelling and simulation 2024. The 2024 European Simulation and Modelling Conference}</div>
<div>editor = {Manuel Graña; J. David Nuñez-Gonzalez}</div>
<div>year = {2024},</div>
<div>month = {10},</div>
<div>pages = {138-144},</div>
<div>publisher = {EUROSIS-ETI},</div>
<div>title = {Media-Text: a Media Industry-Based Dataset for Scene Text Detection}</div>
<div>}</div>
<div>
Attribution 4.0 (CC BY 4.0)https://creativecommons.org/licenses/by/4.0/
License information was derived automatically
This dataset were downloaded from the Open Images V7 repository using the following code through the google colab environment. The classes that make up the dataset are: Boat, Watercraft, Surfboard, Gondola and Jet_ski.
Attribution 4.0 (CC BY 4.0)https://creativecommons.org/licenses/by/4.0/
License information was derived automatically
Public dataset built with images extracted from Open Images Dataset V7 combined with Vehicle Classification V2
MIT Licensehttps://opensource.org/licenses/MIT
License information was derived automatically
A program for managing collections of full spectrum recordings of bats.v6.2.6660 incorporates the import and export of collections of pictures in the image compare window.v6.2.6661 fixes some bugs and speed issues in 6660.v6.2.6680 tries to fix some database updating problems and adds additional debugging in this area.v7.0.6760 - Major improvements and changes.First define the additional shortvut key in Audacity - CTRL-SHIFT-M=Open menu in focussed track. New item in 'View' menu- Analyse and Import, will open a folder of .wav files and sequentially open them in Audacity. When annotated and the label file saved and Audacity closed the next file will be opened. If the label file is not saved then the process stops and will resume on the next invocation of Analyse and Import on that folder. As each file is opened the label track wil be automatically created and named.and the view ill zoom to the first 5 seconds of the .wav track.7.0.6764 also includes a new report format which (for one or more sessions) gives number of minutes in each ten minute window throughout the day in which a species of bat was detected. Rows are given for each species in the recordings. In Excel looks good as a bar chart or a radar chart.7.06789 hopefully fixes the problems when trying to update a database that caused the program to crash on startup if the database did not contain the more recent Version table.7.0.6799 cosmetic changes to use the normal file selection dialog instead of the folder browser dialog, and also when using Analyse and Import, you no longer need to pick a file when selecting the .wav file folder.7.0.6820 Adds session data to all report formats, including pass statistics for all species found in that session.7.0.6844 Adds the ability to add, save, adjust and include in exported images, Fiducial lines. Lines can be added, deleted or adjusted in the image comparison window and are saved to the database when the window is closed. For exported images the lines are permanently overlaid on the image and are no longer adjustable.7.0.6847 Makes slight improvements to the aspect ratio of images in the comparison window and when images are exported the fiducial lines are only included if the FIDS button is deptessed.7.0.6850 Fixes an occasional bug when saving images through Analyse and Import - using filenames in the caption has priority over bat's names. Also improvements in file handling when changing databases - now attempts to recognise if a db is the right type.7.0.6858 Makes some improvements to image handling, including a modification to the database structure to allow long descriptions for images (previously description+caption had to be less than 250 chars) and the ability to copy images within the application (but not to external applications). A single image may now be used simultaneously as a bat image, a call image or a segment image. Changes to it in one location will be reflected in all the other locations. On deletion the link is removed and if there are no remaining links for the image then the image itself will be removed from the database.7.0.6859 has some improvements to the image handling system. In the batReference view the COMP button now adds all bat and call images for all selected bats to the comparison window. Double clicking on a bat adds all bat, call and segment images for all the bats selected to the comparison window.7.0.6860 removed the COMP button from the bat reference view. Double-clicking in this view transfers all images of bat, calls and recordings to the comparison window. Double-clicking in the ListByBats view transfers all recording images but not the bat and call images to the comparison window. Exported images for recordings use the recording filename plus the start offset of the segment as a filename, or alternatively the image caption. 7.0.6866 Improvements to the grids and to grid scaling and movement especially for the sonagram grids.7.0.6876 Added the ability to right-click on a labelled segment in the recordings detail list control, to open that recording in Audacity and scroll to the location of that labelled segment. Only one instance of Audacity may be opened at a time or the scrolling does not work. Also made some improvements to the scrolling behaviour of the recording detail window.Version 7.1 makes significant changes to the way in which the recordingSessions list is displayed. Because this list can get quite large and therefore takes a long time to load, it now loads the data in discrete pages.At the top of the RecordingSessions List is a new navigation bar with a set of buttons and two combo-boxes. The rightmost combobox is used to set the number of items that will be loaded and displayed on a page. The selections are currently 10, 25, 50 and 100. Slower machines may find it advantageous to use smaller page sizes in order to speed up load times and reduce the demand for memory and cpu-time.The other combobox allows the selection of a sort field for the session list. Sessions are displayed in columns in a DataGrid which allows columns to be re-sized, moved and sorted. These functions all now only apply to the subset of data that has been loaded as a page. The Combo-box allows you to sort the full set of data in the database before loading the page. Thus if the combobox is set to sort on DATE with a Page size of 10, then only the 10 earliest (or the 10 latest depending on the direction of sorting) sessions in the database will be loaded. The displayed set of sessions can be sorted on the screen by clicking the column headers but this only changes the order on the screen, it does not load any other sessions from the database.The four buttons can be used to load the next or previous pages or to move to the start or end of the complete database collection. The Next or Previous buttons move the selection by 2/3 of the Page Size so that there will always be some visual overlap between pages.The sort combo-box has two entries for each field, one with a suffix of ^ and one with a suffix of v . These sort the database in Ascending or Descending order. Selecting a sort field will update the display and sort the display entries on the same field, but the sort direction of the displayed items will be whatever was last used. Clicking the column header will change the direction of sort for the displayed items.v7.1.6885 Updates the database to DB version 6.2 by the addition of two link tables between bats and recordings and between bats and sessions. These tables enable much faster access to bat specific data. Also various improvements to improve the speed of loading data when switching to List By Bats view, especially with very large databases.v7.1.6891 Further performance improvements in loading ListByBats and in loading imagesv7.1.6901 Has the ability to perform screen grabs of images without needing an external screen grabber program. Shift-Click on the 'PASTE' button and drag and resize the semi-transparent window to select a screen area, right click in the window to capture that portion of the screen. For details refer to Import/Import Picturesv7.1.6913 Fixed some scaling issues on fiducial lines in the comparison windowv7.1.6915 Bugfix for adjusting fiducial lines - 7.1.6913 removedv7.1.6941 Improvements and adjustments to grid and fiducial line handlingv7.1.6951 Fixes some problems with the Search dialogv7.2.6970 Introduces the ability to replay segments at reduced speed or in heterodyne 'bat detector' mode.v7.2.6971 When opening a recording or segment in Audacity the corresponding .txt file will be opened as a label track. NB this only works if there is only a single copy of Audacity open - subsequent calls with Audacity still open do not open the label track.v7.2.6978 Improvements to Heterodyne playback to use pure sinewave.7.2.6984 Bug fixes and mods to image handling - image captions can now have a region appended in seconds after the file name.---BRM-Aud-Setup_v7_2_7000.exeThis version includes its only private copy of Audacity 2.3.0 portable, which will be placed in the same folder as BRM and has its own pre-configured configuration file appropriate for use with BRM. This will not interfere with any existing installation of Audacity but provides all the Audacity features required by BRM with no further action by the user. BRM will use this version to display .wav files.v7.2.7000 also includes a new report format which is tailored to provide data for the Hertfordshire Mammals, Amphibians and Reptiles survey. It also displays the GPS co-ordinates for the Recording Session as an OS Grid Reference as well as latitude and longitude.v7.2.7010 Speed improvements and bug-fixes to opening and running Audacity through BRM. Audacity portable is now located in C:\audacity-win-portable instead of under the BRM program folder.v7.2.7012 Fixed some bugs in Report generation when producing he Frequency Table. Enabled the AddTag button in the BatReference pane.v7.2.7021 Upgrades the Audacity component to version 2.3.1 and a few minor bug fixes.
Attribution 4.0 (CC BY 4.0)https://creativecommons.org/licenses/by/4.0/
License information was derived automatically
This dataset contains images of an artifical flower platform with different insects sitting on it or flying above it. All images were automatically recorded with the Insect Detect DIY camera trap, a hardware combination of the Luxonis OAK-1, Raspberry Pi Zero 2 W and PiJuice Zero pHAT for automated insect monitoring (bioRxiv preprint).
The following object classes were annotated in this dataset:
View the Health Check for more info on class balance.
You can use this dataset as starting point to train your own insect detection models. Check the model training instructions for more information.
Open source Python scripts to deploy the trained models can be found at the insect-detect GitHub repo.
Attribution 4.0 (CC BY 4.0)https://creativecommons.org/licenses/by/4.0/
License information was derived automatically
Model for detecting dogs, cats and birds. Images and labels taken from Google Open Images Dataset V7
Not seeing a result you expected?
Learn how you can add new datasets to our index.
Dataset Card for Open Images Dataset
This dataset contains images from the Open Images dataset. It includes image URLs, split into training, validation, and test sets.
Dataset Details
Dataset Description
Open Images is a dataset of approximately 9 million URLs to images that have been annotated with image-level labels, bounding boxes, object segmentation masks, and visual relationships.
Curated by: Google LLC License: Images: CC BY 2.0 license… See the full description on the dataset page: https://huggingface.co/datasets/bitmind/open-images-v7.