Data subset loading with regex or string comparison

[bntrtm]

Within the load dialog, users should also have the option of defining what images are loaded based on a provided regex rule (or maybe just based on whether or not a provided string is found in the filename).

This would allow users to more efficiently apply tags to certain images within the dataset that they know share particular similarities, granted that the images do have filenames referencing these similarities.