Multi-scale attention-based instance segmentation for measuring crystals with large size variation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Neubauer, Theresa, Berg, Astrid, Wimmer, Maria, Lenis, Dimitrios, Major, David, Winter, Philip Matthias, De Paolis, Gaia Romana, Novotny, Johannes, Lüftner, Daniel, Reinharter, Katja, Bühler, Katja
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911750544162816
author Neubauer, Theresa
Berg, Astrid
Wimmer, Maria
Lenis, Dimitrios
Major, David
Winter, Philip Matthias
De Paolis, Gaia Romana
Novotny, Johannes
Lüftner, Daniel
Reinharter, Katja
Bühler, Katja
author_facet Neubauer, Theresa
Berg, Astrid
Wimmer, Maria
Lenis, Dimitrios
Major, David
Winter, Philip Matthias
De Paolis, Gaia Romana
Novotny, Johannes
Lüftner, Daniel
Reinharter, Katja
Bühler, Katja
contents Quantitative measurement of crystals in high-resolution images allows for important insights into underlying material characteristics. Deep learning has shown great progress in vision-based automatic crystal size measurement, but current instance segmentation methods reach their limits with images that have large variation in crystal size or hard to detect crystal boundaries. Even small image segmentation errors, such as incorrectly fused or separated segments, can significantly lower the accuracy of the measured results. Instead of improving the existing pixel-wise boundary segmentation methods, we propose to use an instance-based segmentation method, which gives more robust segmentation results to improve measurement accuracy. Our novel method enhances flow maps with a size-aware multi-scale attention module. The attention module adaptively fuses information from multiple scales and focuses on the most relevant scale for each segmented image area. We demonstrate that our proposed attention fusion strategy outperforms state-of-the-art instance and boundary segmentation methods, as well as simple average fusion of multi-scale predictions. We evaluate our method on a refractory raw material dataset of high-resolution images with large variation in crystal size and show that our model can be used to calculate the crystal size more accurately than existing methods.
format Preprint
id arxiv_https___arxiv_org_abs_2401_03939
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Multi-scale attention-based instance segmentation for measuring crystals with large size variation
Neubauer, Theresa
Berg, Astrid
Wimmer, Maria
Lenis, Dimitrios
Major, David
Winter, Philip Matthias
De Paolis, Gaia Romana
Novotny, Johannes
Lüftner, Daniel
Reinharter, Katja
Bühler, Katja
Computer Vision and Pattern Recognition
I.2.10; I.4.6
Quantitative measurement of crystals in high-resolution images allows for important insights into underlying material characteristics. Deep learning has shown great progress in vision-based automatic crystal size measurement, but current instance segmentation methods reach their limits with images that have large variation in crystal size or hard to detect crystal boundaries. Even small image segmentation errors, such as incorrectly fused or separated segments, can significantly lower the accuracy of the measured results. Instead of improving the existing pixel-wise boundary segmentation methods, we propose to use an instance-based segmentation method, which gives more robust segmentation results to improve measurement accuracy. Our novel method enhances flow maps with a size-aware multi-scale attention module. The attention module adaptively fuses information from multiple scales and focuses on the most relevant scale for each segmented image area. We demonstrate that our proposed attention fusion strategy outperforms state-of-the-art instance and boundary segmentation methods, as well as simple average fusion of multi-scale predictions. We evaluate our method on a refractory raw material dataset of high-resolution images with large variation in crystal size and show that our model can be used to calculate the crystal size more accurately than existing methods.
title Multi-scale attention-based instance segmentation for measuring crystals with large size variation
topic Computer Vision and Pattern Recognition
I.2.10; I.4.6
url https://arxiv.org/abs/2401.03939