Jim said that the filter is working exactly as intended, and uses AI learning to identify new shill memes. If some of your meme's aren't working, then they have been flagged either by them being posted from a known shill IP, or from a VPN exit point.
It is what it is anon. It's not in our hands.
>I'm concerned that the filter is doing something simple like a checksum or md5 hash, and then comparing them against known values.
You can do a simple test. If you have an image that is flagged, resize the image, and try again. Resizing will change both of the values that you mentioned. If it's AI examining the image, then it won't.