paused videos get OCR'd into text
- Get Image of Paused Video
- Pull html
- Find relevant video element (with jQuery?)
- If is paused, turn element into image
- An alternative to scraping the html for the video element would be to accept some sort of user input like a user-defined area (a click-and-drag square).
- Apply OCR to image
- Tesseract.js is a possibility
- Though tenserflow would be more versatile and, of course, more fun/cooler
- Output text from image