Data Sets

Normally the GitHub release feature is great for storing pre-processed datasets. In this case, there are some limitations. While a single PNG image is individually small, a directory full of them can be quite large. This is solved with compression, usually in the form of *.tar.gz. The problem is that static has no patterns and does not compress.

Below are the timings for example data sets. These data sets are not intended to be used directly. They are intended to give a visual example of the product. The data sets at the intended quantity (50k) take around 7+ hours each to complete and exceed even GitHub's generous storage. During project planning, make sure to account for the time it takes to generate the full data set.

File Count	Image Size	Type	Color Space	Time to Generate
100	480 x 320	B&W	RGB	00:18
100	480 x 320	B&W	YIQ	00:40
100	480 x 320	Color	RGB	00:30
100	480 x 320	Color	YIQ	00:43