This seems to be a bug in FBReader 0.8.04. I confirmed the problem by looking at
Potter, Beatrix: The Tale of Samuel Whiskers. v1, 16 May 2008, which has the wrong image issue you describe.
One work around would be to convert the MOBI file to HTML+Images. Either MobiPerl or Calibre can do this. I use mobi2oeb from Calibre. The OEB ebook that Calibre produces is not FBReader friendly (the TOC is lost), so there are two options: a) Zip the entire directory and rename from .zip to .oebzip, or b) delete the .oeb file and then zip the entire directory. Under (a) you have an OEB ebook and under (b) you have a single-file HTML (plus images) in a ZIP file. Either one can be read by FBReader, and I confirmed that the HTML ZIP approach worked for the Potter MOBI. If the document did not start out as a MOBI you have even more options, but the basic point is that the bug is for MOBI only so use a different format.
Note that while you have the "exploded" MOBI you could also optimize the size of images if you want to take the time to do so. I don't know of any automatic tools to do this though.