why not make a simple datatable, have you app get the image, decipher it and place it in the table, then push the table to the report. If there were multiple images, you can put the linking information in the table and link the table into the rest of the report.
When you push the data, where it is located is immaterial to CR. Your app has to be able to find it and assemble it, and then hand it to CR to display. Obviously, I am a big proponent of this method. The app knows where the data is, can manipulate it if needed, and finally pass it to the report.
HTH