Zero-Shot Digital Rock Image Segmentation with a Fine-Tuned Segment Anything Model

dc.contributor.authorMa, Zhaoyang
dc.contributor.authorHe, Xupeng
dc.contributor.authorSun, Shuyu
dc.contributor.authorYan, Bicheng
dc.contributor.authorKwak, Hyung
dc.contributor.authorGao, Jun
dc.contributor.departmentPhysical Science and Engineering (PSE) Division
dc.contributor.departmentEarth Science and Engineering Program
dc.contributor.departmentEnergy Resources and Petroleum Engineering Program
dc.contributor.departmentAli I. Al-Naimi Petroleum Engineering Research Center (ANPERC)
dc.contributor.institutionSaudi Aramco EXPEC Advanced Research Center, Reservoir Engineering Technology Division, Pore Scale Physics Focus Area, Bldg. 2291, GA-168
dc.date.accessioned2023-12-04T12:55:59Z
dc.date.available2023-12-04T12:55:59Z
dc.date.issued2023-11-17
dc.description.abstractAccurate image segmentation is crucial in reservoir modelling and material characterization, enhancing oil and gas extraction efficiency through detailed reservoir models. This precision offers insights into rock properties, advancing digital rock physics understanding. However, creating pixel-level annotations for complex CT and SEM rock images is challenging due to their size and low contrast, lengthening analysis time. This has spurred interest in advanced semi-supervised and unsupervised segmentation techniques in digital rock image analysis, promising more efficient, accurate, and less labour-intensive methods. Meta AI's Segment Anything Model (SAM) revolutionized image segmentation in 2023, offering interactive and automated segmentation with zero-shot capabilities, essential for digital rock physics with limited training data and complex image features. Despite its advanced features, SAM struggles with rock CT/SEM images due to their absence in its training set and the low-contrast nature of grayscale images. Our research fine-tunes SAM for rock CT/SEM image segmentation, optimizing parameters and handling large-scale images to improve accuracy. Experiments on rock CT and SEM images show that fine-tuning significantly enhances SAM's performance, enabling high-quality mask generation in digital rock image analysis. Our results demonstrate the feasibility and effectiveness of the fine-tuned SAM model (RockSAM) for rock images, offering segmentation without extensive training or complex labelling.
dc.description.sponsorshipThis work signifies a collaborative endeavor led by Prof. Shuyu Sun and Prof. Bicheng Yan, generously supported by Saudi Aramco (Dr. Hyung Tae Kwak). We would like to acknowledge the Supercomputing Laboratory at King Abdullah University of Science & Technology (KAUST) in Thuwal, Saudi Arabia, for providing the computational resources utilized in this research. We thank anonymous reviewers for their specific comments and instructive suggestions. Thank the Digital Rocks Portal (https://www.digitalrocksportal.org/) for providing the open source data.
dc.eprint.versionPre-print
dc.identifier.arxivid2311.10865
dc.identifier.urihttps://repository.kaust.edu.sa/handle/10754/695915
dc.publisherarXiv
dc.relation.urlhttps://arxiv.org/pdf/2311.10865.pdf
dc.rightsThis is a preprint version of a paper and has not been peer reviewed. Archived with thanks to arXiv.
dc.titleZero-Shot Digital Rock Image Segmentation with a Fine-Tuned Segment Anything Model
dc.typePreprint
display.details.left<span><h5>Type</h5>Preprint<br><br><h5>Authors</h5><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.author=Ma, Zhaoyang,equals">Ma, Zhaoyang</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.author=He, Xupeng,equals">He, Xupeng</a><br><a href="https://repository.kaust.edu.sa/search?query=orcid.id:0000-0002-3078-864X&spc.sf=dc.date.issued&spc.sd=DESC">Sun, Shuyu</a> <a href="https://orcid.org/0000-0002-3078-864X" target="_blank"><img src="https://repository.kaust.edu.sa/server/api/core/bitstreams/82a625b4-ed4b-40c8-865a-d6a5225a26a4/content" width="16" height="16"/></a><br><a href="https://repository.kaust.edu.sa/search?query=orcid.id:0000-0002-3356-7594&spc.sf=dc.date.issued&spc.sd=DESC">Yan, Bicheng</a> <a href="https://orcid.org/0000-0002-3356-7594" target="_blank"><img src="https://repository.kaust.edu.sa/server/api/core/bitstreams/82a625b4-ed4b-40c8-865a-d6a5225a26a4/content" width="16" height="16"/></a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.author=Kwak, Hyung,equals">Kwak, Hyung</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.author=Gao, Jun,equals">Gao, Jun</a><br><br><h5>KAUST Department</h5><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Physical Science and Engineering (PSE) Division,equals">Physical Science and Engineering (PSE) Division</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Earth Science and Engineering Program,equals">Earth Science and Engineering Program</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Energy Resources and Petroleum Engineering Program,equals">Energy Resources and Petroleum Engineering Program</a><br><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.department=Ali I. Al-Naimi Petroleum Engineering Research Center (ANPERC),equals">Ali I. Al-Naimi Petroleum Engineering Research Center (ANPERC)</a><br><br><h5>Date</h5>2023-11-17</span>
display.details.right<span><h5>Abstract</h5>Accurate image segmentation is crucial in reservoir modelling and material characterization, enhancing oil and gas extraction efficiency through detailed reservoir models. This precision offers insights into rock properties, advancing digital rock physics understanding. However, creating pixel-level annotations for complex CT and SEM rock images is challenging due to their size and low contrast, lengthening analysis time. This has spurred interest in advanced semi-supervised and unsupervised segmentation techniques in digital rock image analysis, promising more efficient, accurate, and less labour-intensive methods. Meta AI's Segment Anything Model (SAM) revolutionized image segmentation in 2023, offering interactive and automated segmentation with zero-shot capabilities, essential for digital rock physics with limited training data and complex image features. Despite its advanced features, SAM struggles with rock CT/SEM images due to their absence in its training set and the low-contrast nature of grayscale images. Our research fine-tunes SAM for rock CT/SEM image segmentation, optimizing parameters and handling large-scale images to improve accuracy. Experiments on rock CT and SEM images show that fine-tuning significantly enhances SAM's performance, enabling high-quality mask generation in digital rock image analysis. Our results demonstrate the feasibility and effectiveness of the fine-tuned SAM model (RockSAM) for rock images, offering segmentation without extensive training or complex labelling.<br><br><h5>Acknowledgements</h5>This work signifies a collaborative endeavor led by Prof. Shuyu Sun and Prof. Bicheng Yan, generously supported by Saudi Aramco (Dr. Hyung Tae Kwak). We would like to acknowledge the Supercomputing Laboratory at King Abdullah University of Science & Technology (KAUST) in Thuwal, Saudi Arabia, for providing the computational resources utilized in this research. We thank anonymous reviewers for their specific comments and instructive suggestions. Thank the Digital Rocks Portal (https://www.digitalrocksportal.org/) for providing the open source data.<br><br><h5>Publisher</h5><a href="https://repository.kaust.edu.sa/search?spc.sf=dc.date.issued&spc.sd=DESC&f.publisher=arXiv,equals">arXiv</a><br><br><h5>arXiv</h5><a href="https://arxiv.org/abs/2311.10865">2311.10865</a><br><br><h5>Additional Links</h5>https://arxiv.org/pdf/2311.10865.pdf</span>
kaust.acknowledged.supportUnitSupercomputing Laboratory at King Abdullah University of Science & Technology (KAUST)
kaust.personMa, Zhaoyang
kaust.personSun, Shuyu
kaust.personYan, Bicheng
orcid.authorMa, Zhaoyang
orcid.authorHe, Xupeng
orcid.authorSun, Shuyu::0000-0002-3078-864X
orcid.authorYan, Bicheng::0000-0002-3356-7594
orcid.authorKwak, Hyung
orcid.authorGao, Jun
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
2311.10865.pdf
Size:
3.07 MB
Format:
Adobe Portable Document Format
Description:
Pre-print