Does Tesseract neglect any nontext area in a scanned document?

Question

I'm using Tesseract but I don't know whether it neglects any nontext area and targets text only. Do I have to remove any nontext area as a preprocessing step for better output?

Answer 1

Tesseract has a pretty good algorithm to detect text, but it will eventually give false-positive matches.

Ideally, you would pre-process the image before submitting it to tesseract. Some time ago I engaged in a similar task, so I suggest you take a look at the following material:

• OpenCV C++/Obj-C: Detecting a sheet of paper / Square Detection

• Executing cv::warpPerspective for a fake deskewing on a set of cv::Point

• Rotate cv::Mat using cv::warpAffine offsets destination image

• Affine Transform, Simple Rotation and Scaling or something else entirely?