ZeroChaos

Same problem. Will you be looking for a solution? The plugin is not fully functional, is the author interested in it at all? If not, archive the repository.

Same problem.

Same problem. macOS Sonoma. Are you going to fix it? The issue is open from October 16.

Jitsi uses TFLite for this purpose: https://www.tensorflow.org/lite with the MediaPipe Meet Segmentation model: https://mediapipe.page.link/meet-mc https://ai.googleblog.com/2020/10/background-features-in-google-meet.html and paired with WebAssembly SIMD instructions: https://v8.dev/features/simd

Perhaps something useful can be gleaned from this repository: https://github.com/minhkhue3214/new_virtualbackground

I tried using the @mediapipe/selfie_segmentation library to solve this issue `` I added a blur filter: ``` 'blur': { description: "background blur", f: function (src, width, height, ctx) { if...

If I understand correctly, then: `selfieSegmentation.send({ image: src });` But the image freezes after 8 seconds, so I guess I'm wrong somewhere.

I rewrote the code, but still haven't figured out where I should use asynchronous functions? I now get a static picture immediately when the video conference starts. The background blur...

If I understand correctly, selfieSegmentation.send() will be called asynchronously as part of the setInterval operation. As a result, after clearing the browser cache, the code I gave yesterday works, but...

I have added this functionality to my fork. But from the author the implementation would be better for sure. You can make this feature optional, let administrators decide whether they...