These are things that the converter does not currently handle. Think of these are projects for you to get involved!
Generated lists such as the table of contents (index, list of illustrations, etc) are stored within the ODF file (that is, the generated text is stored file and the application is expected to build the list).
This text should not be extracted for translation, since then a user would see the same text twice (once in the table of contents and again for the heading with the very real likelihood of them being translated differently).
Another problem we have, is that when the heading text is changed, OpenOffice.org won't regenerate auto-generated lists. This is annoying and it would be good to be able to deal with this.