I usually read to improve my vocabularies, like light novels and similars, but lately I’m reading comics online.
I think using OCR is possible to translate dialogues and other texts into images.
I can help implement the solution if you agree with this request.
For example:

Of course, is better if we can do a better job than this.