Watershed uses only one CPU and runs very slow on large volumes

[qin-yu]

I'm running PlantSeg for FOR2581 but found watershed step uses only one CPU. I'll try to improve it but please stop me if it's not possible @lorenzocerrone

[lorenzocerrone]

Parallelizing the distance transform might be possible in Vigra, but I think the watershed is single-core only

[qin-yu]

Parallelizing the distance transform might be possible in Vigra, but I think the watershed is single-core only

Cry :' (

It took > 35 minutes to run watershed on a clear nuclear boundary pmap of size [473, 1712, 1712]! I have 63 cores idling

INFO:  PlantSeg Napari info - 15:15:55 04.06.2024 - Watershed Segmentation: Widget Watershed Segmentation computation started
Missing chrome or resource URL: resource://gre/modules/UpdateListener.jsm
Missing chrome or resource URL: resource://gre/modules/UpdateListener.sys.mjs
INFO:  PlantSeg Napari info - 15:51:27 04.06.2024 - Watershed Segmentation: Widget Watershed Segmentation computation complete in 2132.12s