Looking at and thinking about datasets to use in some forthcoming ML projects I started thinking about some good uses for training models on images. Since my grandmother died of lung cancer that was missed by doctors that eventually spread to her brain, I went looking for large datasets of CT scans for cancer. There is a plethora of data sets from CT scans, the Cancer Imaging Archive is linked above and contains many datasets for all types of cancer. Of note is the last line of the above image.

“Two deep learning researchers used the images and the corresponding annotation files to train several well-known detection models which resulted in a maximum a posteriori probability (MAP) of around 0.87 on the validation set.”

Its interesting to see this note about deep learning being used in cancer diagnosis.