Amirkabir University of Technology AUT Journal of Electrical Engineering 2588-2910 58 4 2026 10 01 SaveDoc: A Robust Technique for Degradation Classification on Ancient Document Images 681 694 6118 10.22060/eej.2026.24823.5773 EN Khairun Saddami Department of Electrical and Computer Engineering, Faculty of Engineering, Universitas Syiah Kuala, Banda Aceh, Indonesia 0000-0001-8344-339X Rizka Ramadhana Department of Electrical and Computer Engineering, Faculty of Engineering, Universitas Syiah Kuala, Banda Aceh, Indonesia Fathurrahman Fathurrahman Eletrical Engineering Department, College of Engineering and Physics, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia Department of Electrical and Computer Engineering, Faculty of Engineering, Universitas Syiah Kuala, Banda Aceh, Indonesia Khairul Munadi Department of Electrical and Computer Engineering, Faculty of Engineering, Universitas Syiah Kuala, Banda Aceh, Indonesia 0000-0002-7507-9476 Fitri Arnia Department of Electrical and Computer Engineering, Faculty of Engineering, Universitas Syiah Kuala, Banda Aceh, Indonesia 0000-0001-6020-1275 Journal Article 2025 09 30 The analysis of degraded ancient documents remains challenging because multiple degradation patterns may appear within a single document image and reduce document quality. In this paper, we propose a combination of convolutional neural networks and extreme learning machine for patch-level degradation classification in ancient document images. The dataset was constructed from 100 source document images and cropped into 700 labeled patches. Four pre-trained architectures, namely VGG19, ResNet101, MobileNetV2, and EfficientNetB0, were used as feature extractors. The number of hidden nodes in the ELM classifier was tuned, and two activation functions, namely radial basis function and sigmoid, were evaluated. The experimental results show that EfficientNetB0 features combined with the extreme learning machine classifier achieved the best performance, with a testing accuracy of 98.21% ± 2.08%. This combination also produced faster inference time compared with the other evaluated models. These results indicate that the EfficientNetB0-extreme learning machine combination is effective for patch-level degradation classification in ancient document images. https://eej.aut.ac.ir/article_6118_1755c118e8859eb000eb6eca25369407.pdf