Improved scanning process.
Preprocces to recognize features of forms such as boxes with or without checks and the boxes around fields.
I scan image documents to be able to search for technical reviews or to export them to excel, access, or word. Current situation: On forms empty or full check boxes are frequently recognized as text and the text is wrong and takes a long time to correct. The same occurs with text near boundaries around form fields. If form type features such as check boxes and field boundaries can be recognized by a scanning algorithm very different from the small box scan process used to recognize text. If done many non text pixels can be removed from the pixels for the text scan eliminating many errors and speeding the text scan process. The orie3ntation information from the form features may also contribute to make text scaning more accurate. There is perhaps in the current text scan process many lines of code devoted to difficulty caused by non-text form artifacts. Knowledge of form boundaries can also contribute to the recognition process. Some non text blobs and imperfections located too near the boundaries ought to contribute to the presumption the it is non-text. Horizontal field lines can contribute to the likelihood estimate that some unknown blob is text if that possible text is at a 15 degree angle or more to the box orientation. I am attaching an image showing my version of Adobe Acrobat Pro DC. I do wuiswh to end saying this is a fantaqstic prodiuct already and has improved a great deal since I first started using Acrobat.
Thanks for the suggestion. We will look into this. We will update you once there is an update.