Still the image data may not be convertible to UTF-8, its just a sequence of bytes, not any encoding.
UTF-8 no, but many encodings are actually "a sequence of bytes", which is the reason why choosing the right one when opening is so hard (basically, it's a guessing game if you don't know already the answer). *Any* stream of bytes its gonna be valid in one of the ISO-8859-* encodings, the only reason we don't accept them is that we refuse NUL bytes.