how do I import a video?

[Raj2032]

I want to instead of trying to scan a still image I want to instead import a video file and detect all the text inside it and print it to the console maybe I could try and use opencv if this is possible? just not quite too sure how I would get it integrated into ocrs?

[robertknight]

To process a video you will need to:

1. Find a way to extract frames from the video as an image. I am not that familiar with the state of video processing in Rust, but a wrapper for GStreamer, FFmpeg or OpenCV might be a starting point (eg. the VideoCapture APIs)
2. Convert the image to 8-bit RGB, if not already in that format
3. Use ocrs::ImageSource::from_bytes to prepare the image for Ocrs to process.
4. Once you have an ImageSource, you can then feed it into text recognition following the steps in the example.

To speed up extraction, you will probably want to sample only a subset of the frames. For example, extract only 1 in every N frames.

[Raj2032]

ok thanks mate :)