backscrub
backscrub copied to clipboard
floe
Virtual Video Device for Background Replacement with Deep Semantic Segmentation
To avoid clipping the image partway it would be nice if the image may be split into two (or more) overlapping areas that are fed to the NN and recombined...
The full meet model often leaves small holes in my face. It could be worthwile to add an optional hole filling step.
It'd be nice to feed an already-recorded video through this to remove the background.
I finally reached to work with the animated.gif file. I had thought that ffmpeg will be used, this was not the case. After installing the gstreamer plugin, I was able...
Crash
If we call backscrub with the debug mode set to 2 and the size of the camera stream or the size of the output stream is to small, eg. 160x90...
If we don't have the same aspect ratio for the camera and the video output the result is not very pretty. For this case, we can calculate the region we...
I have performed tests with an HDMI-grabber attached to a USB-3 input. The video source was from a video camera. The frame rate for the v4l2 device was 50 FPS,...
The arguments passed to backscrub are checked by the function strncmp(). strncmp() should be replaced with strcmp(). Wrong arguments are not always recognized and may produce erroneous messages. For example:...
System: Fedora 36 Backscrub: version heads/main-0-g4f70bc1 (Tensorflow: build tags/v2.8.0-0-g3f878cff5b6, run-time 2.8.0) opencv 4.5.5 Launching backscrub with: ``` backscrub -c /devvideo0 -v /dev/video3 -m body-pix-float-050-8.tflite ``` crash at libbackscrub.cc line 368...