Citation

Lew, Kai Liang and Sim, Kok Swee and Tan, Shing Chiang and Gan, Kok Beng (2026) Deep Learning-Based SEM Image Noise Level Classification. International Journal on Advanced Science, Engineering and Information Technology, 16 (2). pp. 672-684. ISSN 2088-5334

Text Kai+Liang+21058-AAP.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Scanning electron microscopy (SEM) allows high-resolution visualization of surface structures at the nanoscale in materials science, the chip industry, and the biomedical field. However, Gaussian noise can degrade image quality and complicate the interpretation analyze of SEM images. Researchers commonly use classical image filtering methods, such as median, Wiener, and Gaussian filters, to address this issue. However, challenges remain when selecting an optimal filter and its parameters on different datasets. This study aims to develop seven modified versions of the MobileNetV3 architecture for classifying Gaussian noise variances ranging from 0.001 to 0.01 in SEM images. These models were trained and evaluated on a dataset of 3,630 SEM images. The dataset was split into training, validation, and test sets with a 60:20:20 ratio. All images were resized to 128 x 128 pixels. MobileNetV3 Large with the fifth modification, trained with a batch size of 32, achieved strong performance, including 98.89% accuracy, precision, recall, and F1 score, indicating improved noise classification. These results show that the MobileNetV3 Large (fifth version) model classifies noise levels in SEM images. Real SEM noise is often mixed with other noise sources and may not follow a Gaussian distribution, so future work should validate the approach on real noisy acquisitions and extend it to additional noise types. In conclusion, this study demonstrates the effectiveness of deep learning techniques for noise classification in microscopy images.

Item Type:	Article
Uncontrolled Keywords:	Scanning electron microscopy, deep learning
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Engineering and Technology (FET) Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	30 Jun 2026 04:34
Last Modified:	30 Jun 2026 04:34
URII:	http://shdl.mmu.edu.my/id/eprint/16133

Downloads

Edit (login required)