img-hash VS opencv-rust

I absolutely loved all the crates available that made my life very simple in many cases. I used (among others) the slint ui framework, kamadak-exif, img_hash, fast_image_resize and rawloader.

I'm using default img_hash algorithm - Gradient. It is also possible to use with this library different algorithms - https://github.com/abonander/img_hash/blob/master/src/alg/mod.rs#L20-L81

https://github.com/twistedfall/opencv-rust makes it evident that they do try to support using a prebuild system-opencv. Maybe it is not correctly recognized?

opencv

The de facto solution for smart image processing is to use a computer vision library, such as opencv, which has experimental rust bindings: https://github.com/twistedfall/opencv-rust

There is an OpenCV bindings crate. This example might be what you're after https://github.com/twistedfall/opencv-rust/blob/master/examples/video_capture.rs

For example, this is the build script of the opencv bindings: https://github.com/twistedfall/opencv-rust/blob/master/build.rs

I’m trying to get around a known build issue—it would be great for it to be fixed upstream, but I don’t have any control over it except for setting OUT_DIR during its compilation.

I've been using https://github.com/twistedfall/opencv-rust a lot lately and it's really fantastic, they've done a great job matching up with how C++ works, so there are definitely good examples out there.

I never tried it, but the OpenCV bindings might work.

I thought it was going to be unsafe because of how Mat is implemented (more information here). ... I tested it and it should be fine memory-wise. My thought was that the code below was going to be allowed to compile.