The original string is not in any way part of the image that's generated. The fuzzer notices that the initial codepath being triggered with the "hello" file would be different if the first byte were 0xff, instead of 0x68. So it changes the file and tries it. The 'h' has gone - it wouldn't matter what it was originally, the fuzzer would always have chosen 0xff.

All the fuzzer is doing is exploring the possible codepaths through the application trying to exercise all the code; many of the codepaths end up with the executable outputting an error message and terminating. Some maybe put it into an infinite loop. Some end up with it completing a JPEG data parse and terminating - so in amongst all the possible paths it explores, of course it will eventually seek out input sequences which bring that about.