The University of Southampton
University of Southampton Institutional Repository

Segment anything model for medical images?

Segment anything model for medical images?
Segment anything model for medical images?

The Segment Anything Model (SAM) is the first foundation model for general image segmentation. It has achieved impressive results on various natural image segmentation tasks. However, medical image segmentation (MIS) is more challenging because of the complex modalities, fine anatomical structures, uncertain and complex object boundaries, and wide-range object scales. To fully validate SAM's performance on medical data, we collected and sorted 53 open-source datasets and built a large medical segmentation dataset with 18 modalities, 84 objects, 125 object-modality paired targets, 1050K 2D images, and 6033K masks. We comprehensively analyzed different models and strategies on the so-called COSMOS 1050K dataset. Our findings mainly include the following: (1) SAM showed remarkable performance in some specific objects but was unstable, imperfect, or even totally failed in other situations. (2) SAM with the large ViT-H showed better overall performance than that with the small ViT-B. (3) SAM performed better with manual hints, especially box, than the Everything mode. (4) SAM could help human annotation with high labeling quality and less time. (5) SAM was sensitive to the randomness in the center point and tight box prompts, and may suffer from a serious performance drop. (6) SAM performed better than interactive methods with one or a few points, but will be outpaced as the number of points increases. (7) SAM's performance correlated to different factors, including boundary complexity, intensity differences, etc. (8) Finetuning the SAM on specific medical tasks could improve its average DICE performance by 4.39% and 6.68% for ViT-B and ViT-H, respectively. Codes and models are available at: https://github.com/yuhoo0302/Segment-Anything-Model-for-Medical-Images. We hope that this comprehensive report can help researchers explore the potential of SAM applications in MIS, and guide how to appropriately use and develop SAM.

Medical image segmentation, Medical object perception, Segment anything model
1361-8415
Huang, Yuhao
0ea59da6-b6ad-48b9-8fb6-1cb471a4ca09
Yang, Xin
472eb5fb-3f3a-4a11-a95e-90d7dc79586e
Liu, Lian
7b0a8c8b-0c18-4bc7-9f60-902533008cb8
Zhou, Han
528f9270-a93a-40b5-98ee-66794057b0dd
Chang, Ao
d8a33dee-aba2-4f95-91e2-5b051dde739d
Zhou, Xinrui
a9d04545-0e72-4d15-95d9-a0d06f1a8200
Chen, Rusi
977e35e3-0d60-48ef-9108-73291b11b61e
Yu, Junxuan
4c87f5af-a2e2-416d-a76b-ea9852b13c05
Chen, Jiongquan
b5799746-57fc-45c0-bbe6-fcf489eceec8
Chen, Chaoyu
dc322474-4bb5-4879-8bde-36fca518e78a
Liu, Sijing
a29c959d-a4cb-4250-b645-7b8805500a83
Chi, Haozhe
b80b586a-23be-47f8-9b07-b9b7560620d1
Hu, Xindi
b3d6983d-63e9-4e2f-9b59-c5a3d4bca9b7
Yue, Kejuan
975570ad-f412-40af-ad86-ac45db5b9a40
Li, Lei
2da88502-0bd8-4e6b-8f7d-0c01a48b399e
Grau, Vicente
c6992187-7840-45ce-93a0-d54d06031c7d
Fan, Deng Ping
1a1f8010-8f1b-44a6-8f8d-e09cbeff81c0
Dong, Fajin
db94a719-6902-40e5-b219-9b7a6d045013
Ni, Dong
7cd4fbeb-f641-43a9-bf5e-0549ffdeaa43
et al.
Huang, Yuhao
0ea59da6-b6ad-48b9-8fb6-1cb471a4ca09
Yang, Xin
472eb5fb-3f3a-4a11-a95e-90d7dc79586e
Liu, Lian
7b0a8c8b-0c18-4bc7-9f60-902533008cb8
Zhou, Han
528f9270-a93a-40b5-98ee-66794057b0dd
Chang, Ao
d8a33dee-aba2-4f95-91e2-5b051dde739d
Zhou, Xinrui
a9d04545-0e72-4d15-95d9-a0d06f1a8200
Chen, Rusi
977e35e3-0d60-48ef-9108-73291b11b61e
Yu, Junxuan
4c87f5af-a2e2-416d-a76b-ea9852b13c05
Chen, Jiongquan
b5799746-57fc-45c0-bbe6-fcf489eceec8
Chen, Chaoyu
dc322474-4bb5-4879-8bde-36fca518e78a
Liu, Sijing
a29c959d-a4cb-4250-b645-7b8805500a83
Chi, Haozhe
b80b586a-23be-47f8-9b07-b9b7560620d1
Hu, Xindi
b3d6983d-63e9-4e2f-9b59-c5a3d4bca9b7
Yue, Kejuan
975570ad-f412-40af-ad86-ac45db5b9a40
Li, Lei
2da88502-0bd8-4e6b-8f7d-0c01a48b399e
Grau, Vicente
c6992187-7840-45ce-93a0-d54d06031c7d
Fan, Deng Ping
1a1f8010-8f1b-44a6-8f8d-e09cbeff81c0
Dong, Fajin
db94a719-6902-40e5-b219-9b7a6d045013
Ni, Dong
7cd4fbeb-f641-43a9-bf5e-0549ffdeaa43

Huang, Yuhao, Yang, Xin and Liu, Lian , et al. (2023) Segment anything model for medical images? Medical Image Analysis, 92, [103061]. (doi:10.1016/j.media.2023.103061).

Record type: Article

Abstract

The Segment Anything Model (SAM) is the first foundation model for general image segmentation. It has achieved impressive results on various natural image segmentation tasks. However, medical image segmentation (MIS) is more challenging because of the complex modalities, fine anatomical structures, uncertain and complex object boundaries, and wide-range object scales. To fully validate SAM's performance on medical data, we collected and sorted 53 open-source datasets and built a large medical segmentation dataset with 18 modalities, 84 objects, 125 object-modality paired targets, 1050K 2D images, and 6033K masks. We comprehensively analyzed different models and strategies on the so-called COSMOS 1050K dataset. Our findings mainly include the following: (1) SAM showed remarkable performance in some specific objects but was unstable, imperfect, or even totally failed in other situations. (2) SAM with the large ViT-H showed better overall performance than that with the small ViT-B. (3) SAM performed better with manual hints, especially box, than the Everything mode. (4) SAM could help human annotation with high labeling quality and less time. (5) SAM was sensitive to the randomness in the center point and tight box prompts, and may suffer from a serious performance drop. (6) SAM performed better than interactive methods with one or a few points, but will be outpaced as the number of points increases. (7) SAM's performance correlated to different factors, including boundary complexity, intensity differences, etc. (8) Finetuning the SAM on specific medical tasks could improve its average DICE performance by 4.39% and 6.68% for ViT-B and ViT-H, respectively. Codes and models are available at: https://github.com/yuhoo0302/Segment-Anything-Model-for-Medical-Images. We hope that this comprehensive report can help researchers explore the potential of SAM applications in MIS, and guide how to appropriately use and develop SAM.

This record has no associated files available for download.

More information

Accepted/In Press date: 5 December 2023
e-pub ahead of print date: 7 December 2023
Published date: 11 December 2023
Additional Information: Publisher Copyright: © 2023 Elsevier B.V.
Keywords: Medical image segmentation, Medical object perception, Segment anything model

Identifiers

Local EPrints ID: 488803
URI: http://eprints.soton.ac.uk/id/eprint/488803
ISSN: 1361-8415
PURE UUID: dedfa6c7-e597-4f5f-ad7b-806ee503cf1b
ORCID for Lei Li: ORCID iD orcid.org/0000-0003-1281-6472

Catalogue record

Date deposited: 05 Apr 2024 16:44
Last modified: 10 Apr 2024 02:14

Export record

Altmetrics

Contributors

Author: Yuhao Huang
Author: Xin Yang
Author: Lian Liu
Author: Han Zhou
Author: Ao Chang
Author: Xinrui Zhou
Author: Rusi Chen
Author: Junxuan Yu
Author: Jiongquan Chen
Author: Chaoyu Chen
Author: Sijing Liu
Author: Haozhe Chi
Author: Xindi Hu
Author: Kejuan Yue
Author: Lei Li ORCID iD
Author: Vicente Grau
Author: Deng Ping Fan
Author: Fajin Dong
Author: Dong Ni
Corporate Author: et al.

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×