I use Google Keep’s OCR (which is pretty dang good by the way) to pull unknown words from games and content I’m playing on my PC. It does an incredible job at recognizing characters from pretty much anything, so I struggle to use anything else for ripping text from a game or show (other than when I can just use subs directly, or Textractor/Agent).
Currently the process is, 1. Take a Screenshot. 2. Paste that Screenshot into Google Keep. 3. Wait a few seconds. 4. Click a dropdown and hit, Grab text.
From there I can use Yomichan and it’s honestly not too bad when your just grabbing a few things here and there, but when you do this with over hundreds if not thousands of screenshots it can start to get tedious. I would really like something that would let me just screengrab the area the text is in, run it through keeps OCR and dumps it into my clipboard. I don’t mind if I have to wait the time for the processing, it’s just the clicking through each button/menu that gets tedious.
I’ve tried other OCR’s like Manga OCR, Tessaract via Game2Text and some others and while they do work fine sometimes, Keep has a near 99% accuracy rate for me so I’ll keep putting up with this if I must, but just hoping someone out there knows some neat API trick or something I don’t.
by Aboreric