The idea hit me when I was reading the subtitles to my youngest son on a Sunday morning while we were watching a movie. At the same moment I thought of my mother, who has about 5% of eyesight left. She would also benefit from an app like this.
We are Dutch and although he is quite good in understanding English already, understanding everything that is said is too difficult still. His ability to read the subtitles fast enough is also not yet developed enough.
My idea was:
- Make an app which takes a photo every second of a (TV) screen.
- Limit the area of where the photo is taken to where the subtitles appear.
- Feed the photo to an OCR service.
- Compare the text that is returned with the last text.
- If the text is different, let the app read the text (TTS).
I've written a test app which can take a photo every second and feed it to an OCR service.
At the moment I am still developing my JAVA skills (I am a PHP developer) , so developing the app goes step by step.