How to use deepstream to got a frames batched data

52 Views Asked by At

I would like to use deepstream batch infer, where the model inputs are consecutive video frames, e.g. batch * 3 * 16 * 224 * 224. where 3 is the number of channels, 16 is the number of frames, and 224 is the length and width. I found that gst-nvstreammux can pack video frames from multiple sources into a batch. But this does not guarantee that the number of frames from each source in the output is at least 16. Is there any plugin that takes as input video frames from n sources and outputs m consecutive 16 frames, m <= n,as a result I got a m * 3 * 16 * 224 * 224 data.

0

There are 0 best solutions below