Hi Team,
I’m working with a Luxonis segmentation light model and I can run inference on a folder of images. The model produces the usual highlighted segmentation output, but I need to access the raw segmentation masks (e.g., binary 0/1 mask for tuna meat vs background).
The goal is to use the masks to crop out the tuna meat itself from the images, since that’s what we’ll be processing in later stages of the pipeline.
Is it possible to retrieve the mask during inference? If so, what would be the recommended way to access it? I'm not sure if I am pissing anything obvious here.
Thanks in advance for any guidance!
