BTW, if all you need from PIL is image size info, you might want to check out the python code in KindleUnpack that handles generation of covers called mobi_cover.py.
It can be found here:
https://github.com/kevinhendricks/Ki.../mobi_cover.py
There is a bug fixed image type identifier and the original image size routine that PIL used to create their version that was written by Draco.
If you use that version, you may not need PIL at all.