Multiclass Evaluation of Vision Transformers for Industrial Welding Defect Detection

Мультиклассовая оценка трансформеров зрения для обнаружения дефектов промышленной сварки
Antonio Contreras Ortiz, Ricardo Rioda Santiago, Daniel E. Hernandez, Miguel Angel Lopez-Montiel
2025-02-28

Aluminum 5083 TIG datasetVision Transformersindustrial quality inspectionmulticlass classificationwelding defect detection
Automating industrial processes, particularly quality inspection, is a key objective in manufacturing. While welding tasks are frequently automated, inspection processes remain largely manual. Advances in computer vision and AI, especially ViTs, now enable more effective defect detection and classification, offering opportunities to automate these workflows. This study evaluates ViTs for identifying defects in aluminum welding using the Aluminum 5083 TIG dataset. The analysis spans binary classification (detecting defects) and multiclass categorization (Good Weld, Burn Through, Contamination, Lack of Fusion, Misalignment, and Lack of Penetration). ViTs achieved 98% to 99% accuracy across both tasks, significantly outperforming prior models such as dense and CNNs, which struggled to surpass 80% accuracy in binary and 70% in multiclass tasks. These results, achieved with datasets of 2400 to 8000 images, highlight ViTs’ efficiency even with limited data. The findings underline the potential of ViTs to enhance manufacturing inspection processes by enabling faster, more reliable, and cost-effective automated solutions, reducing reliance on manual inspection methods.
1
The multiclass task distinguishes six categories: Good Weld, Burn Through, Contamination, Lack of Fusion, Misalignment, and Lack of Penetration.
2
The results indicate that Vision Transformers can achieve high inspection accuracy with relatively limited datasets containing 2,400–8,000 images.
3
The study evaluates Vision Transformers for binary and multiclass aluminum welding defect detection using the Aluminum 5083 TIG dataset.
4
Vision Transformers achieved 98%–99% accuracy across both binary and multiclass classification tasks.
5
Vision Transformers substantially outperformed dense and convolutional neural networks, which remained below 80% accuracy in binary and 70% in multiclass classification.

aluminum 5083 TIG welds

Vision Transformer performance for binary welding-defect detection and multiclass defect classification

Publication Details
Publication Date
2025-02-28
Journal
Publisher
ISSN
Cited by
21
Access Type
Author Information
Authors
Antonio Contreras Ortiz
Ricardo Rioda Santiago
Daniel E. Hernandez
Miguel Angel Lopez-Montiel
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%