How do I use it for multiple lines of handwritten text?

#5
by abhikatta - opened

So, I want to use the model on a chunk of handwritten text, I tried splitting the image using OpenCV, but it didn't work as expected, either I get duplicates of each line, or I get nothing. If anyone could help me with this, I would greatly appreciate it.

Same problem

you may first use text detectors like CRAFT to cut out the text regions, then send those images to TrOCR to get recognitions.

use EasyOCR to get location of text, then use those location to take cutout from image and perform TrOCR on it.

Sign up or log in to comment