Reconstruction of concrete morphology using deep learning

In this contribution, the concrete morphology is reconstructed with a simple algorithm selecting a pixel value based on the small set of surrounding pixels. A deep neural network (DNN) is used as a classifier, and the authors focus on studying different DNN architectures. The performance of the pro...

Full description

Saved in:
Bibliographic Details
Main Authors: Ondřej Šperl, Jan Sýkora
Format: Article
Language:English
Published: Czech Technical University in Prague 2024-11-01
Series:Acta Polytechnica CTU Proceedings
Subjects:
Online Access:https://ojs.cvut.cz/ojs/index.php/APP/article/view/10214
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850057666227863552
author Ondřej Šperl
Jan Sýkora
author_facet Ondřej Šperl
Jan Sýkora
author_sort Ondřej Šperl
collection DOAJ
description In this contribution, the concrete morphology is reconstructed with a simple algorithm selecting a pixel value based on the small set of surrounding pixels. A deep neural network (DNN) is used as a classifier, and the authors focus on studying different DNN architectures. The performance of the proposed algorithm is evaluated on several statistical descriptors and the grain size distribution curve.
format Article
id doaj-art-b9844e99cced46fb9bf50101e2c11c59
institution DOAJ
issn 2336-5382
language English
publishDate 2024-11-01
publisher Czech Technical University in Prague
record_format Article
series Acta Polytechnica CTU Proceedings
spelling doaj-art-b9844e99cced46fb9bf50101e2c11c592025-08-20T02:51:23ZengCzech Technical University in PragueActa Polytechnica CTU Proceedings2336-53822024-11-014910.14311/APP.2024.49.0085Reconstruction of concrete morphology using deep learningOndřej Šperl0Jan Sýkora1Czech Technical University in Prague, Faculty of Civil Engineering, Department of Mechanics, Thákurova 7, 160 00 Prague, Czech RepublicCzech Technical University in Prague, Faculty of Civil Engineering, Department of Mechanics, Thákurova 7, 160 00 Prague, Czech Republic In this contribution, the concrete morphology is reconstructed with a simple algorithm selecting a pixel value based on the small set of surrounding pixels. A deep neural network (DNN) is used as a classifier, and the authors focus on studying different DNN architectures. The performance of the proposed algorithm is evaluated on several statistical descriptors and the grain size distribution curve. https://ojs.cvut.cz/ojs/index.php/APP/article/view/10214reconstructionconcretedeep learningconvolutional neural network
spellingShingle Ondřej Šperl
Jan Sýkora
Reconstruction of concrete morphology using deep learning
Acta Polytechnica CTU Proceedings
reconstruction
concrete
deep learning
convolutional neural network
title Reconstruction of concrete morphology using deep learning
title_full Reconstruction of concrete morphology using deep learning
title_fullStr Reconstruction of concrete morphology using deep learning
title_full_unstemmed Reconstruction of concrete morphology using deep learning
title_short Reconstruction of concrete morphology using deep learning
title_sort reconstruction of concrete morphology using deep learning
topic reconstruction
concrete
deep learning
convolutional neural network
url https://ojs.cvut.cz/ojs/index.php/APP/article/view/10214
work_keys_str_mv AT ondrejsperl reconstructionofconcretemorphologyusingdeeplearning
AT jansykora reconstructionofconcretemorphologyusingdeeplearning