Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ Recolector de Cienci...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Computers in Industry
Article . 2021 . Peer-reviewed
License: CC BY
Data sources: Crossref
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Computers in Industry
Article
License: CC BY
Data sources: UnpayWall
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
http://dx.doi.org/10.1016/j.co...
Article
License: Elsevier TDM
Data sources: Sygma
https://dx.doi.org/10.60692/m1...
Other literature type . 2021
Data sources: Datacite
https://dx.doi.org/10.60692/0a...
Other literature type . 2021
Data sources: Datacite
DBLP
Article . 2022
Data sources: DBLP
Computers in Industry
Article . 2021 . Peer-reviewed
versions View all 8 versions
addClaim

Ensemble of convolutional neural networks based on an evolutionary algorithm applied to an industrial welding process

مجموعة من الشبكات العصبية الالتفافية بناءً على خوارزمية تطورية مطبقة على عملية لحام صناعية
Authors: Yarens J. Cruz; Marcelino Rivas; Ramón Quiza; Alberto Villalonga; Rodolfo E. Haber; Gerardo Beruvides;

Ensemble of convolutional neural networks based on an evolutionary algorithm applied to an industrial welding process

Abstract

Cet article présente une approche de classification d'images basée sur un ensemble de réseaux de neurones convolutionnels et l'application à une étude de cas réelle d'un procédé de soudage industriel. L'ensemble se compose de cinq réseaux de neurones convolutionnels, dont les sorties sont combinées par le biais d'une politique de vote. Afin de sélectionner les paramètres de réseau appropriés (c'est-à-dire le nombre de couches convolutives et d'hyperparamètres de couches) et la politique de vote, un processus de recherche efficace a été effectué à l'aide d'un algorithme évolutif. La méthode proposée est appliquée et validée dans une étude de cas axée sur la détection d'un désalignement des tôles à assembler par un procédé de soudage à l'arc immergé. Après avoir sélectionné la configuration la plus pratique, l'ensemble surpasse les sept autres stratégies considérées dans une comparaison dans plusieurs métriques, tout en maintenant un coût de calcul adéquat.

Este artículo presenta un enfoque para la clasificación de imágenes basado en un conjunto de redes neuronales convolucionales y la aplicación a un estudio de caso real de un proceso de soldadura industrial. El conjunto consta de cinco redes neuronales convolucionales, cuyos resultados se combinan a través de una política de votación. Para seleccionar los parámetros de red apropiados (es decir, el número de capas convolucionales y los hiperparámetros de las capas) y la política de votación, se llevó a cabo un proceso de búsqueda eficiente utilizando un algoritmo evolutivo. El método propuesto se aplica y valida en un estudio de caso centrado en la detección de desalineación de chapas metálicas a unir mediante proceso de soldadura por arco sumergido. Después de seleccionar la configuración más conveniente, el conjunto supera a otras siete estrategias consideradas en una comparación en varias métricas, al tiempo que mantiene un coste computacional adecuado.

This paper presents an approach for image classification based on an ensemble of convolutional neural networks and the application to a real case study of an industrial welding process. The ensemble consists of five convolutional neural networks, whose outputs are combined through a voting policy. In order to select appropriate network parameters (i.e., the number of convolutional layers and layers hyperparameters) and voting policy, an efficient search process was carried out by using an evolutionary algorithm. The proposed method is applied and validated in a case study focused on detecting misalignment of metal sheets to be joined through submerged arc welding process. After selecting the most convenient setup, the ensemble outperforms other seven strategies considered in a comparison in several metrics, while maintaining an adequate computational cost.

تقدم هذه الورقة مقاربة لتصنيف الصور بناءً على مجموعة من الشبكات العصبية الالتفافية وتطبيقها على دراسة حالة حقيقية لعملية اللحام الصناعي. تتكون المجموعة من خمس شبكات عصبية التفافية، يتم دمج مخرجاتها من خلال سياسة التصويت. من أجل اختيار معلمات الشبكة المناسبة (أي عدد الطبقات الالتفافية والمعلمات الفائقة للطبقات) وسياسة التصويت، تم إجراء عملية بحث فعالة باستخدام خوارزمية تطورية. يتم تطبيق الطريقة المقترحة والتحقق من صحتها في دراسة حالة تركز على اكتشاف عدم محاذاة الصفائح المعدنية التي سيتم ربطها من خلال عملية لحام القوس المغمور. بعد اختيار الإعداد الأكثر ملاءمة، تتفوق المجموعة على الاستراتيجيات السبع الأخرى التي يتم أخذها في الاعتبار في المقارنة في العديد من المقاييس، مع الحفاظ على تكلفة حسابية كافية.

Keywords

Artificial neural network, Artificial intelligence, Image classification, FOS: Political science, Machine Vision, FOS: Mechanical engineering, Convolutional neural network, FOS: Law, Pattern recognition (psychology), Fabric Defect Detection in Industrial Applications, Industrial and Manufacturing Engineering, Defect Detection, Engineering, Ensemble learning, Machine learning, Welding, Political science, Laser Welding, Evolutionary parameters, Hyperparameter, Mechanical Engineering, Politics, Computer science, Mechanical engineering, Ensemble of models, Process (computing), Algorithm, Operating system, Mechanics of Materials, Welding Techniques and Residual Stresses, Physical Sciences, Convolutional neural networks, Surface Defect Detection, Voting, Applications of Infrared Thermography in Non-Destructive Testing, Law

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    45
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Top 1%
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Top 10%
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Top 1%
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
45
Top 1%
Top 10%
Top 1%
Green
hybrid