Reformulate instance label transforms

constantinpape / torch-em

Deep-learning based semantic and instance segmentation for 3D Electron Microscopy and other bioimage analysis problems based on pytorch.

MIT License

76 stars 22 forks source link

Open anwai98 opened 1 month ago

anwai98 commented 1 month ago

EDIT: We update the current version of instance label transforms update function:

i.e. the function used to get label transforms to extract desired labels (eg, binary, boundaries and offsets supported in most LM datasets): formalize the arguments for this in a bit more unified way, such that they receive a string input and track argument values via kwargs.

constantinpape commented 1 month ago

I agree in principle, but let's discuss the details before starting with an implementation.