thank you for this amazing project. I used it as a base to further improve the recognition algorithm as follows:
Custom compiled opencv.js that only contains the functions we need => reduced file size from ~9 MB to about 1.8 MB.
For paper extraction: Use OpenCV's grabCut algorithm to remove the background for higher precision when using the extractPaper function (=> not used in realtime highlighting as the algorithm needs a few seconds to complete).
Use morphologyEx to remove noise and text inside the documents.
Use Canny edge detection to detect the boundaries of objects.
Add cv.dilate to account for small inaccuracies in the Canny edge detection.
Add a white border around the image to be able to detect documents not completely contained inside the canvas.
Removed thresholding since its replaced by Canny edge and morphologyEx
I would just like to share it, maybe it is useful for other people as well. Maybe you can test it as well and consider adding some changes to your project.
Hi,
thank you for this amazing project. I used it as a base to further improve the recognition algorithm as follows:
I would just like to share it, maybe it is useful for other people as well. Maybe you can test it as well and consider adding some changes to your project.
jscanify_modified.zip