Adaptive fuzzy weighted median filter for microcalcifications detection in digital breast tomosynthesis images

• Published in: Indonesian Journal of Electrical Engineering and Computer Science
• Main Author: Saifudin S.A.; Sulaiman S.N.; Osman M.K.; Isa I.S.; Karim N.K.A.; Harron N.A.
• Format: Article
• Language: English
• Published: Institute of Advanced Engineering and Science 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85186682899&doi=10.11591%2fijeecs.v34.i1.pp197-209&partnerID=40&md5=332d48a94b6a6ab29688f978918dfab1

Breast cancer is a global leading cause of female mortality. Digital breast tomosynthesis (DBT) is pivotal for early breast cancer detection, with microcalcifications serving as crucial indicators. However, the movement of the DBT machine introduces blurry artefacts, potentially impacting accurate diagnosis. This study addresses this challenge by proposing an adaptive fuzzy weighted median filter (AFWMF) to enhance DBT images and aid microcalcification diagnosis. AFWMF automatically determines optimal parameters based on input images, outperforming conventional methods with a threshold range (C) from peak to end of switching. Quantitative assessment reveals peak signal to noise ratio (PSNR), and mean absolute error (MAE) values of 96.2267 and 0.0000636, respectively, demonstrating a significant improvement in microcalcification detection. This study contributes an effective and adaptive enhancement technique for DBT images, promising better breast cancer diagnosis, particularly in microcalcification scenarios. © 2024 Institute of Advanced Engineering and Science. All rights reserved.