Open LauraErhard opened 5 years ago
There is a dedicated backend method, where the new box coordinates are sent, and the backend does its magic generating new boxes. That is the reason why the boxes snap to certain positions according to the OCR results (even if they are overlapping).
We noticed that for reference lists where the space between references is very narrow, some boxes can't get "fixed". This results to this: