Free Datasets
Free Receipt OCR Dataset
The dataset consists of 192 images with a total of 3,839 bounding boxes, where each box has a different class.
Extreme Keypoints Dataset
This dataset consists of 480 images collected from copyright-free sources where each image contains a person in extreme or contorted pose.
Car Parts and Car Damages Dataset
The dataset consists of a total of 1812 images, fully annotated with polygons either of car parts and car damages.
Teeth Segmentation Dataset
The dataset consists of 598 dental images with 15,318 polygons representing each tooth segmented with a different class.
Recycling Dataset
The dataset consists of 3000 images, with "litter" as the main class and further tags such as Material, Object and Brand.
Arabic Documents OCR Dataset
The dataset consists of 10K images, that are further classified into 12 classes namely Invoices, Books, labels, etc.
Supermarket Shelves Dataset
This dataset consists of 45 images with a total of 11,743 bounding boxes with the classes Product and Price.
Plant Segmentation Dataset
The dataset contains 144 images of plant seedlings shot at different time intervals within the span of 2 months
Daily Objects Around the World Detection
27k images with bounding box and tag annotations released in partnership with Gapminder's Dollar Street project
Medical Mask and Accessory Detection
6k images for detecting medical masks and other accessories, published as a contribution to the fight against COVID-19
Semantic Segmentation of Aerial Imagery
72 images for drone and satellite vision prepared in partnership with the MBR Space Centre in Dubai, UAE
Dollar Street Dataset
The dataset includes 27519 images grouped into 138 folders. The images are sorted in 3 types: Abstract, Places and bounding box annotations.
Need a dataset labeled pro bono?
We are available to label datasets for social impact projects for free. If you are working on such a project and you need help labeling your dataset, get in touch with us.