It needs a lot of work, and requires ImageMagick to run, there are no safeguards built in yet, it also assumes sed and vim are installed.
Usage involves changing to the directory and running the script with a basename of an html index to be generated, and a txt file which consists of checksum width height filesize and filename.
I'd like to use this to maintain a basic database to prevent archiving duplicates (I've got a lot of work to do). When vim pops up, delete any lines of files that you wish to save. If anyone has a better suggestion on how to handle that with less user interaction, or at least make it more pleasant than firing up an editor, it would be appreciated.
Please note that this is no longer being developed. I currently use fdupes which compiled fine under macports, is available in homebrew, and is probably available in you favorite Linux Distribution's repository.