Could this be an issue with the browser and the way it handles JPG files? Do you have any applications defined for JPG files in Firefox?
Changing the downloaded file extension to .txt and opening in Notepad will allow seeing the file header which may give a clue as to what the actual file type is. It sounds like you are saving a link to the original file or a it is html that has been given the wrong suffix.
I have been able to duplicate the same issue with Firefox 34 and Windchill 10.1 and 10.2. As mentioned, it works as expected with IE and Chrome and also with changing the Attachments Download Mechanism to Java or DTI.
If the file is downloaded with FF34, it has the right file name, but contains the HTML to dispay the image and not the data for the image. That is why all the viewers tell me its an invalid image file.
This one will need a little more homework...
This is just some weird behavior on Firefox. It is not reproducible on IE or Chrome. PTC's QA team has reported SPR 2253339 to address this problem.