iNaturalist is another good place to find photographs of biological organisms. You can easily filter for images with specified levels of copyright. This link will get you to a page with observations of genus Haliaeetus in the public domain (copyrights waived by the photographer under the CC0 designation):
In addition to simply searching sources like Google Images or Getty Images, the Global Biodiversity Information Facility (GBIF) has a gallery feature that allows you to search for images of an organism. Note that (1) images are user contributed, therefore the licensing is determined by the contributor, and (2) are not hosted by GBIF directly. Make sure ...
Multimodal Brain Tumor Segmentation Challenge 2019
Imaging Data Description
All BraTS multimodal scans are available as NIfTI files (.nii.gz) and describe a) native (T1) and b) post-contrast T1-weighted (T1Gd), c) T2-weighted (T2), and d) T2 Fluid Attenuated Inversion Recovery (T2-FLAIR) volumes, and were acquired with different clinical protocols and ...
A Dataset for Sky Segmentation
The Sky dataset contains a collection of 60 images with ground truth for sky segmentation. It was based on the Caltech Airplanes Side dataset by R. Fergus 15/02/03. Those images from the dataset which contained sky regions were selected and ground truths were created for them. The original dataset image names were ...
Instagram is a start. Just scrape and use opencv or similar to identify faces. Or, manually sort faces.
Instagram is a simple http site that can be crawled recursively. It's a matter of counting the images so tweak it and find a good balance. Nobody is going to take 1000 pictures of the same face so dont expect perfection.
Even github enterprise has a 100 GB hard limit. I'd consider Amazon S3 as an alternative, especially because they price based on options for how often the data is accessed and also provides very granular user access facilities. There's a cost calculator
You could start with a benchmark using the MNIST dataset. There are many tutorials and articles written on training a CNN for digit recognition. You might also want to consider preprocessing your color-blindness charts input data (e.g. converting to grayscale, enhancing digit contrast).
Here are some links from government agencies:
The trick is to use str inside the contains.
This query gives the expected result:
select ?item ?img
hint:Query hint:maxParallel 400 .
hint:Query hint:chunkSize 8000 .
?item wdt:P18 ?img .
filter contains(str(?img), "Montage")
Another way to achieve this is to set hint:Query hint:regexMatchNonString true and then use a normal regex.
Check out the ICDAR 2019 Robust Reading Challenge on Scanned Receipts OCR and Information Extraction: https://rrc.cvc.uab.es/?ch=13
You have to make a free account and then you can access the dataset, approximately 1000 images of scanned receipts.
As I already answered on a later question, a list of Medical Imaging datasets can be found on the medical-imaging-datasets repository. Also, it could be interesting to follow both Stephen R. Aylward's list of repositories and the SICAS Medical Image Repository.