Post

To achieve these improvements, we finetune SAM on a large LM dataset (ca. 50k images, 1.2M annotated cells) and on a large EM dataset (ca. 5k images, 90k annotated nuclei and mitos). We also add a new decoder for instance segmentation, which provides efficient automatic segmentation.

Our models form the basis of micro_sam, our napari plugin for interactive and automatic segmentation. It can segment data in 2D, 3D and across time. You can find all the details at github.com/computationa...

After a long journey 😂💗💗💗

Things have also improved a lot along the journey :). (But I still wouldn't have minded one fewer round of revisions ;))

This is so cool. Do you think it would work to segment and quantify 3D STED images?