Couldn't screen readers apply unicode normalization based some heuristics, like seeing the continuous presence of those special bold/italic characters? To impro

Couldn't screen readers apply unicode normalization based some heuristics, like seeing the continuous presence of those special bold/italic characters? To improve accuracy, it can even check if the normalized text resembles to some English words or phrases or not.