New tutorial series is out!
In it I show how you can :
- Scan your own books using video, not pictures, so it’s easier to record.
- Automatically detect page turns to find “keyframes” from the page turning video
- Extract the text using OCR (using AWS Textract and Microsoft Read)
- Check OCR accuracy using Project Gutenberg
- And generate audio files from text using the Microsoft Text to Speech API
I hope it’s interesting to you. Check out the Github and Youtube Channel
Why I Made It
I really don’t like how various ebook stores operate. Essentially you license the content and don’t own it. If they ever shut down you may be royally screwed. Also your access is general controlled by using approved interfaces.
Buying a book means you own it forever and you don’t have to worry if Kindle or whoever shuts down one day. And be free from censorship being introduced after the fact.
Also, many books that are freely available on Project Gutenberg are sold on book marketplaces when it could be free. I hope this helps people trying to work on similar projects changing how you interact with books.
There’s this cool App called Matter (getmatter.app) that does a lot of this on their platform but still I think there should be a open source solution.