Meta’s Segment Anything Model 2 was announced last week. Are there any thoughts about its applicability to 3D images?

Maybe this is of interest: [image] Oxford researchers unveil MedSAM-2, an AI that could change how doctors... Researchers have developed a new AI model called MedSAM-2 that significantly improves the segmentation of 2D and 3D medical images. With just a single manual annotation, it can …

Also this: [2408.03322] Segment Anything in Medical Images and Videos: Benchmark and Deployment With a Slicer module! GitHub - bowang-lab/MedSAMSlicer at SAM2 [SAM2 3D Slicer Plugin: Usage]

I was aware of MedSAMSlicer, I didn’t notice they have a SAM2 branch :smile: Thank you for pointing it out.

They uploaded their preprint five days after the SAM 2 preprint …

Many people celebrate performance of SAM when they see some demo videos, but they don’t realize that they are looking at trivial segmentation tasks (e.g., kidney segmentation on a single CT slice where no similar structures are nearby) and they don’t notice when the segmentation spectacularly fails …

TotalSegmentator, MONAIAuto3DSeg, DentalSegmentator, etc. Ground truth for training them needs to come from somewhere. SAM is a great interactive segmentation tool.

[image] dzenanz: SAM is a great interactive segmentation tool Yes, and this is its main limitation: it requires interaction. If the user needs to interact with the image for several (potentially tens of) seconds then the value is questionable, because the clinician can make the standard, well…

SAM is foreign to me, I won’t comment about SAM. [image] lassoan: reduces the clinician’s effort This means that clinicians want already processed data; i.e, results only. For decades, radiologists have provided that and are still doing that. ‘AI’ tools are trained with normal data and cert…

[image] lassoan: SAM/SAM2/MedSAM/SAM-Med3D/etc. might find their niche where they can be useful in medical imaging, maybe in the future they can even carve out a larger area where they are successful. I agree with this comment. My experience with SAM has been less than ideal for real segmenta…

For awareness, here’s a new preprint by NVIDIA discussing this topic: A Short Review and Evaluation of SAM2’s Performance in 3D CT Image Segmentation | Abstract (arxiv.org)

Does SAM2 in Slicer only run tiny model? [图片]

SAM 2 with video support

Development

Xiaojie_Guo (Xiaojie Guo) October 11, 2024, 2:40am 12

Does SAM2 in Slicer only run tiny model?

Topic		Replies	Views
Integration of “Segment Anything Model”? Development	23	4165	June 17, 2023
Struggling to integrate SAM model to 3D slicer, need help Development	6	916	June 26, 2023
SegmentWithSAM module not working Support segmentation	4	172	October 23, 2025
Run SAMM (Segment Any Medical Model) in 3D Slicer Support segmentation , python	13	1115	October 24, 2023
SAM 2 from Meta (Segment Anything Model) Support	0	191	August 9, 2024

SAM 2 with video support

Related topics