I developed Android document scanner using OpenCV edge detection functions which does not work all the time as expected and requires a special background in order to work properly like dark plain background. Also document color might affect the accuracy of document edge detection. All such requirements are real frustration to the app user.
Searching the Internet I learned that all professional mobile document scanners on google play use machine learning like Tensorflow or any other DML, in order for the scanner app to work under all circumstances with great accuracy.
I went through TensorFlow examples like image classification and image segmentation. I do not know if any of these could be used for document edge detection. Can anybody help me with using TensorFlow for developing android document edge detection app (document scanner).
Thanks
