Is there a way to increase inference time for both segmentation and background matting.The code is taking single image and processing on it.Can we increase batch size and processes multiple frames at a time.
Yeah, you can change the code to use multiple batches. You can use a smaller segmentation network like mobilnetv2. Also check this issue #41 for tips on making this real-time.
Is there a way to increase inference time for both segmentation and background matting.The code is taking single image and processing on it.Can we increase batch size and processes multiple frames at a time.