Then how hard would it be to use some pattern-based image recognition to detect this label? Not very hard, I have a friend that does something similar at work.
You’re doubting the ease of implementation, but it’s really not hard. It’ll have to be a fairly predictable pattern from YouTube’s side, which is one of the easiest to accurately detect.
Nevermind that though, it’ll never be embebbed in the video stream, because it has to be accessible as well as readable. It’s impossible to guarantee it to be readable without actually rendering the text in whatever client it’s being viewed in. Imagine a 240p video, the text would have to take up half the screen to be readable with that low resolution.
Then how hard would it be to use some pattern-based image recognition to detect this label? Not very hard, I have a friend that does something similar at work.
Lol, good luck with that!
You’re doubting the ease of implementation, but it’s really not hard. It’ll have to be a fairly predictable pattern from YouTube’s side, which is one of the easiest to accurately detect.
Nevermind that though, it’ll never be embebbed in the video stream, because it has to be accessible as well as readable. It’s impossible to guarantee it to be readable without actually rendering the text in whatever client it’s being viewed in. Imagine a 240p video, the text would have to take up half the screen to be readable with that low resolution.
There’s literally no problem to render a text in a video stream.