ngalstyan4's profile - activity

Hi!

I am trying to identify a data table from a photo of a document. The photo is taken from a skewed angle so some preprocessing is necessary before feeding to an OCR engine like tesseract.

After some preprocessing Here is what I have. It is clearly visible where the table is. I however am having difficulties removing the noise from the sides so I can identify the minAreaRect around the table then do a 4-point transform. Can you give me any ideas how to go about doing it?