Cover Image

Timber defect identification: Enhanced classification with residual networks

This study investigates the potential enhancement of classification accuracy in timber defect identification through the utilization of deep learning, specifically residual networks. By exploring the refinement of these networks via increased depth and multi-level feature incorporation, the goal is...

Full description

Saved in:
Bibliographic Details
Main Authors: Teo, Hong Chun, Hashim, Ummi Rabaah, Ahmad, Sabrina
Format: Article
Language:English
Published: Science and Information Organization 2024
Online Access:http://eprints.utem.edu.my/id/eprint/27549/2/0167809052024165055.PDF
http://eprints.utem.edu.my/id/eprint/27549/
https://thesai.org/Downloads/Volume15No4/Paper_68-Timber_Defect_Identification.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.utem.eprints.27549
record_format eprints
spelling my.utem.eprints.275492024-07-24T16:26:57Z http://eprints.utem.edu.my/id/eprint/27549/ Timber defect identification: Enhanced classification with residual networks Teo, Hong Chun Hashim, Ummi Rabaah Ahmad, Sabrina This study investigates the potential enhancement of classification accuracy in timber defect identification through the utilization of deep learning, specifically residual networks. By exploring the refinement of these networks via increased depth and multi-level feature incorporation, the goal is to develop a framework capable of distinguishing various defect classes. A sequence of ablation experiments was conducted, comparing our proposed framework’s performance (R1, R2 and R3) with the original ResNet50 architecture. Furthermore, the framework’s classification accuracy was evaluated across different timber species and statistical analyses such as independent t-tests and one-way ANOVA tests were conducted to identify the significant differences. Results showed that while the R1 architecture demonstrated slight improvement over ResNet50, particularly with the addition of an extra layer ("ConvG"), the latter still maintained superior overall performance in defect identification. Similarly, the R2 architecture, despite achieving notable accuracy improvements, slightly lagged behind R1. Integration of fully pre-activation activation functions in the R3 architecture yielded significant enhancements, with a 14.18% increase in classification accuracy compared to ResNet50. The R3 architecture showcased commendable defect identification performance across various timber species, though with slightly lower accuracy in Rubberwood. Nonetheless, its performance surpassed both ResNet50 and other proposed architectures, suggesting its suitability for timber defect identification. Statistical analysis confirmed the superiority of the R3 architecture across multiple timber species and this underscores the significance of integrating network depth and fully pre-activation activation functions in improving classification performance. In conclusion, while the wood industry has made strides towards automation in timber grading, significant challenges remain. Overcoming these challenges will require innovative approaches and advancements in image processing and artificial intelligence to realize the full potential of automated grading systems. Science and Information Organization 2024 Article PeerReviewed text en http://eprints.utem.edu.my/id/eprint/27549/2/0167809052024165055.PDF Teo, Hong Chun and Hashim, Ummi Rabaah and Ahmad, Sabrina (2024) Timber defect identification: Enhanced classification with residual networks. International Journal of Advanced Computer Science and Applications, 15 (4). pp. 665-671. ISSN 2158-107X https://thesai.org/Downloads/Volume15No4/Paper_68-Timber_Defect_Identification.pdf 10.14569/IJACSA.2024.0150468
institution Universiti Teknikal Malaysia Melaka
building UTEM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Teknikal Malaysia Melaka
content_source UTEM Institutional Repository
url_provider http://eprints.utem.edu.my/
language English
description This study investigates the potential enhancement of classification accuracy in timber defect identification through the utilization of deep learning, specifically residual networks. By exploring the refinement of these networks via increased depth and multi-level feature incorporation, the goal is to develop a framework capable of distinguishing various defect classes. A sequence of ablation experiments was conducted, comparing our proposed framework’s performance (R1, R2 and R3) with the original ResNet50 architecture. Furthermore, the framework’s classification accuracy was evaluated across different timber species and statistical analyses such as independent t-tests and one-way ANOVA tests were conducted to identify the significant differences. Results showed that while the R1 architecture demonstrated slight improvement over ResNet50, particularly with the addition of an extra layer ("ConvG"), the latter still maintained superior overall performance in defect identification. Similarly, the R2 architecture, despite achieving notable accuracy improvements, slightly lagged behind R1. Integration of fully pre-activation activation functions in the R3 architecture yielded significant enhancements, with a 14.18% increase in classification accuracy compared to ResNet50. The R3 architecture showcased commendable defect identification performance across various timber species, though with slightly lower accuracy in Rubberwood. Nonetheless, its performance surpassed both ResNet50 and other proposed architectures, suggesting its suitability for timber defect identification. Statistical analysis confirmed the superiority of the R3 architecture across multiple timber species and this underscores the significance of integrating network depth and fully pre-activation activation functions in improving classification performance. In conclusion, while the wood industry has made strides towards automation in timber grading, significant challenges remain. Overcoming these challenges will require innovative approaches and advancements in image processing and artificial intelligence to realize the full potential of automated grading systems.
format Article
author Teo, Hong Chun
Hashim, Ummi Rabaah
Ahmad, Sabrina
spellingShingle Teo, Hong Chun
Hashim, Ummi Rabaah
Ahmad, Sabrina
Timber defect identification: Enhanced classification with residual networks
author_facet Teo, Hong Chun
Hashim, Ummi Rabaah
Ahmad, Sabrina
author_sort Teo, Hong Chun
title Timber defect identification: Enhanced classification with residual networks
title_short Timber defect identification: Enhanced classification with residual networks
title_full Timber defect identification: Enhanced classification with residual networks
title_fullStr Timber defect identification: Enhanced classification with residual networks
title_full_unstemmed Timber defect identification: Enhanced classification with residual networks
title_sort timber defect identification: enhanced classification with residual networks
publisher Science and Information Organization
publishDate 2024
url http://eprints.utem.edu.my/id/eprint/27549/2/0167809052024165055.PDF
http://eprints.utem.edu.my/id/eprint/27549/
https://thesai.org/Downloads/Volume15No4/Paper_68-Timber_Defect_Identification.pdf
_version_ 1806430045005152256
score 13.211869

Similar Items