NucleusInstanceSegmentor.predict() has no "patch" mode

[alevangel]

• TIA Toolbox version: 1.5.1
• Python version: 3.8
• Operating System: Windows

What I Did

I would like to do some test with patch mode prediction for Nucleus Instance Segmentation, but I see that is still not implemented?

Here is stated that can be used:

tiatoolbox/tiatoolbox/models/engine/semantic_segmentor.py

Lines 1279 to 1285 in 0002963

	Args:
	imgs (list, ndarray):
	List of inputs to process. When using `"patch"` mode,
	the input must be either a list of images, a list of
	image file paths or a numpy array of an image list. When
	using `"tile"` or `"wsi"` mode, the input must be a list
	of file paths.

But here we assume that could be only 'tile' or 'wsi':

tiatoolbox/tiatoolbox/models/engine/semantic_segmentor.py

Lines 1347 to 1349 in 0002963

	if mode not in ["wsi", "tile"]:
	msg = f"{mode} is not a valid mode. Use either `tile` or `wsi`."
	raise ValueError(msg)

[shaneahmed]

We are working on improving the structure of engines in #578

[shaneahmed]

We are working on improving the structure of engines in #578

The patch mode should still work as NucleusInstanceSegmentor is inheriting from SemanticSegmentor [doc]