Hi,
I’m trying to implement a flow that “distort” time of a video stream.
As a basic test of this feature, I’d like to implement a workflow where the video stream is read from a .gxf file, slowed down by duplicating each image and then written to a .gxf file.
For this, I would think of implementing a custom operator taking one image as input and emiting two signals as output, each one having a copy of the image.
Is this even possible with Holoscan SDK ?
As an alternative, I thought of having a custom operator taking one image as input and emiting one signal as output containing an array of images. I know this is possible, however I don’t see how to write the next steps of the workflow to take each image in sequence to finally write one .gxf file with each image in sequence.
Is this something that could be implemented using Holoscan SDK ? If yes, could you provide some advice on how to reach this goal.
thanks,
-Sebastien
PS: the end goal of course is to process the duplicated images individually and differently