Performing Step-wise Prediction on a Large Image in SAHI

[DhanrajJain77]

I am working with SAHI for image processing and need to perform step-wise prediction on a large image. However, some pixels in the image contain no data value, and I want to skip those slices during inference.

Currently, when I use the predict or get_sliced_prediction method in SAHI, it considers all slices including those with no data value. Is there a way to remove these slices before performing inference?

Additionally, I'm wondering if it's possible to slice the image first, remove the unwanted slices, and then perform the inference on the remaining slices directly. Later, I would like to combine the results with post-processing. I aim to reduce the time. Any guidance or suggestions on how to achieve this in SAHI would be greatly appreciated.