I hope I am wrong, but I think that we currently only have boxes with pixel coordinates for Takahashi's transliteration (thanks to Job's Voynichese.com); while we currently have uncertain spaces only for the ZL transliteration.
I guess that most ZL .-spaces will be spaces for Takahashi too, but probably about half the uncertain ,-spaces will not be spaces for Takahashi (so we don't have coordinates for them). Then the two transliterations will also be different in other ways (as the ar/or uncertainty you mention): mapping one into the other would probably require complex and partially unreliable heuristics.
Rene certainly is the most qualified to comment about the state of the art and the possible approaches.
I guess that most ZL .-spaces will be spaces for Takahashi too, but probably about half the uncertain ,-spaces will not be spaces for Takahashi (so we don't have coordinates for them). Then the two transliterations will also be different in other ways (as the ar/or uncertainty you mention): mapping one into the other would probably require complex and partially unreliable heuristics.
Rene certainly is the most qualified to comment about the state of the art and the possible approaches.