I have an inference pipeline that operates on RGB frames from a RealSense camera. The pipeline structure looks something like this:
(realsense) -> (background subtraction) -> (object detection) -> (object tracking) -> (object classification)
I'm using the segmentation filter for background subtraction, which adds an alpha channel to the RGB frames and embeds the mask into it (the alpha channel). My object detector works on the RGB channels and ignores the alpha channel, so it ignores the background subtraction. I need to apply the alpha channel as a mask to the other channels before passing the frames into the inference segment of the pipeline. I could write a custom filter but would prefer to use one of the existing plugins, are there any existing plugins that handle this? I can drop the alpha channel with videoconvert, but that does me no good since the mask is embedded in the alpha channel.
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel
Locked
