Convolutional Neural Networks Based White Blood Cell Classification Method from Digital Holographic Microscope Images Sayisal Holografik Mikroskop G r nt lerinden Evrisimli Sinir Aglari Tabanli Beyaz Kan H creleri Siniflandirma Y ntemi

Uytan S., Necip S., ESMER G. B.

33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025, İstanbul, Turkey, 25 - 28 June 2025, (Full Text) identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/siu66497.2025.11111769
  • City: İstanbul
  • Country: Turkey
  • Keywords: Classification, Convolutional Neural Network, Digital Holography Microscopy, In-line Hologram, White Blood Cells
  • Marmara University Affiliated: Yes

Abstract

In the presented work, an innovative method is proposed for the classification of white blood cells (WBCs) using digital holography microscopy (DHM) images. The developed approach uses a peripheral blood cell dataset to automatically and accurately classify WBCs. The classification process is performed based on the calculated features and these features are obtained by processing the DHM images via a convolutional neural network (CNN) architecture. By integrating DHM with image processing, we propose an efficient and highly accurate WBC analysis method. The performance of the proposed method is evaluated according to the F1 score and precision metrics. The proposed method gives a test accuracy of 96.45%, an F1 score of 96.44%, and a precision metric of 96.44% for the ResNeXt model, which yielded the best results among the tested architectures.