2021
DOI: 10.3390/sym13071176
|View full text |Cite
|
Sign up to set email alerts
|

Synthetic Data Generation for Steel Defect Detection and Classification Using Deep Learning

Abstract: The paper presents a methodology for training neural networks for vision tasks on synthesized data on the example of steel defect recognition in automated production control systems. The article describes the process of dataset procedural generation of steel slab defects with a symmetrical distribution. The results of training two neural networks Unet and Xception on a generated data grid and testing them on real data are presented. The performance of these neural networks was assessed using real data from the… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
5

Citation Types

2
37
0
1

Year Published

2021
2021
2023
2023

Publication Types

Select...
6
4

Relationship

1
9

Authors

Journals

citations
Cited by 98 publications
(40 citation statements)
references
References 24 publications
2
37
0
1
Order By: Relevance
“…The following traditional methods are most commonly used in industry to refine the structure of such metals: heat treatment based on phase transformations, cold pressure treatment of metals followed by heat treatment, often in the form of recrystallization annealing, and thermomechanical treatment. Such conventional methods allow to grind the microstructure to ultrafine grains (d ≈ 1–10 μm) [ 4 , 5 , 6 , 7 , 8 ]. In recent years, many articles have been written showing that there are severe plastic deformation (SPD) methods that can remove this limitation and grind the microstructure to 0.1 μm and below directly during deformation [ 9 , 10 , 11 , 12 , 13 , 14 , 15 ].…”
Section: Introductionmentioning
confidence: 99%
“…The following traditional methods are most commonly used in industry to refine the structure of such metals: heat treatment based on phase transformations, cold pressure treatment of metals followed by heat treatment, often in the form of recrystallization annealing, and thermomechanical treatment. Such conventional methods allow to grind the microstructure to ultrafine grains (d ≈ 1–10 μm) [ 4 , 5 , 6 , 7 , 8 ]. In recent years, many articles have been written showing that there are severe plastic deformation (SPD) methods that can remove this limitation and grind the microstructure to 0.1 μm and below directly during deformation [ 9 , 10 , 11 , 12 , 13 , 14 , 15 ].…”
Section: Introductionmentioning
confidence: 99%
“…Expansive clay soils spread over large areas all over the world. The properties of these problematic soils at shallow depth show volume change from variations in soil moisture, due to seepage of water into the soils during rains [ 18 , 19 ]. These soils contain the clay mineral montmorillonite, which causes swelling and shrinking in the ground due to increasing and decreasing water content [ 20 , 21 ].…”
Section: Introductionmentioning
confidence: 99%
“…In [ 22 , 23 , 24 ], it was proposed to use synthesized calcium silicate hydrates (CSH) to increase the durability of lime coatings. A lime composition has been developed for finishing and restoring the walls of buildings in the form of a dry mixture containing a filler based on CSH and allowing to obtain mortar mixtures with a water-retaining capacity of 98–99%, drying time of 15–20 min, pot life 1–1.5 h. Coatings based on the proposed dry mortar (DM) are characterized by a vapor permeability coefficient of 0.05 mg/m·h·Pa, adhesion strength of 0.6–0.9 MPa, compressive strength of 3–4 MPa.…”
Section: Introductionmentioning
confidence: 99%