Achieving 100% Accuracy in Steel Surface Defect Classification Using a Custom High-Performance CNN

<title>Abstract</title> <p>Steel surface defect detection is a critical task in industrial quality control. In this study, we propose a custom high-performance Convolutional Neural Network (CNN) architecture specifically designed for classifying six types of steel surface defects using the NEU-CLS dataset. Our model achieves 100% accuracy, 100% precision, 100% recall, and 100% F1-score, establishing a new global state-of-the-art on the NEU-CLS benchmark. This result surpasses all previously reported methods and demonstrates that a well-engineered, task-specific architecture can outperform general-purpose models in specialized industrial inspection tasks. The code and trained models are publicly available for full reproducibility</p>