English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Efficient and accurate extraction of microstructures in micrographs ofmaterials is essential in process optimization and the exploration ofstructure-property relationships. Deep learning-based image segmentationtechniques that rely on manual annotation are laborious and time-consuming andhardly meet the demand for model transferability and generalization on varioussource images. Segment Anything Model (SAM), a large visual model with powerfuldeep feature representation and zero-shot generalization capabilities, hasprovided new solutions for image segmentation. In this paper, we proposeMatSAM, a general and efficient microstructure extraction solution based onSAM. A simple yet effective point-based prompt generation strategy is designed,grounded on the distribution and shape of microstructures. Specifically, in anunsupervised and training-free way, it adaptively generates prompt points fordifferent microscopy images, fuses the centroid points of the coarselyextracted region of interest (ROI) and native grid points, and integratescorresponding post-processing operations for quantitative characterization ofmicrostructures of materials. For common microstructures including grainboundary and multiple phases, MatSAM achieves superior zero-shot segmentationperformance to conventional rule-based methods and is even preferable tosupervised learning methods evaluated on 16 microscopy datasets whosemicrographs are imaged by the optical microscope (OM) and scanning electronmicroscope (SEM). Especially, on 4 public datasets, MatSAM shows unexpectedcompetitive segmentation performance against their specialist models. Webelieve that, without the need for human labeling, MatSAM can significantlyreduce the cost of quantitative characterization and statistical analysis ofextensive microstructures of materials, and thus accelerate the design of newmaterials.

MatSAM: Efficient Extraction of Microstructures of Materials via Visual Large Model