Made the one implemented ilastik object classification model accomodate only the ilastik workflow that requires pixel prediction maps

Pixel classification in 3D works, but pixel maps are not aligned with original image, suspect X-Y flipped

# Conflicts:
#	imagej/infer_ilastik_by_api.py
#	tests/test_ilastik.py

Implemented workflow and endpoint for two-stage (px -> ob) classification in ilastik; tests pass and image data is correct