Research Experiences for Undergraduates in Computational Methods with Applications in Materials Science
MSU Affiliation
College of Arts and Sciences; Department of Mathematics and Statistics; Center for Computational Sciences
Research Mentor
Tung-Lung Wu
Creation Date
7-25-2025
Abstract
Nanoproducts are a growing sector due to their unique properties and wide range of applications across industries. However, nanomaterial production is a complex process that requires a high degree of precision, making it challenging to ensure consistent quality at a large scale. Minor defects can significantly alter their functional properties and overall performance, making accurate detection of defects crucial for maintaining and controlling nanomaterial properties. To address these challenges, this project applies scan statistics to detect localized defects in scanning election microscope images of nanofibrous materials. We implemented both square and circular scanning windows of varying sizes to identify clusters in the images. By generating null distributions through permutation testing, we assessed the statistical significance of detected regions, leading to reliable identification of anomalies. For each image, we identified the most effective scan window by selecting the size that minimized the p-value, allowing us to adapt the detection process to the unique spatial features of each image and effectively address different types of anomalies. Overall, this approach provides a robust and adaptable method for automated anomaly detection, with the potential to enhance quality control in nanomaterial manufacturing.
Presentation Date
Summer 7-31-2025
Keywords
scan statistics, nanomaterials, anomaly detection
Recommended Citation
Anderson, Kaitlyn; Duwage, Asanka; and Wu, Tung-Lung, "Anomaly Detection in Material Images Using Scan Statistics" (2025). Research Experiences for Undergraduates in Computational Methods with Applications in Materials Science. 16.
https://scholarsjunction.msstate.edu/ccs-reu/16
