Impact of Normalization Techniques on Mitosis Detection in H&E Stained Histopathology Images HandE Boyali Histopatoloji Görüntülerinde Normalizasyon Tekniklerinin Mitoz Tespitine Etkisi


Gören A. A., Hançer E., Samet R., Nemati N.

34th Signal Processing and Communications Applications Conference, SIU 2026, İstanbul, Türkiye, 7 - 10 Temmuz 2026, (Tam Metin Bildiri)

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/siu71813.2026.11636553
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: H&E staining, histopathology, Mitosis detection, normalization
  • Ankara Üniversitesi Adresli: Evet

Özet

The aim of this study is to investigate the impact of stain normalization approaches on mitosis detection in histopathology images. Variations in staining and imaging conditions can significantly affect model performance, yet normalization is often applied without a clear understanding of its contribution. To address this, seven widely used stain normalization methods are applied to the ICPR12 and ICPR14 datasets, and their effects are evaluated using YOLOv5 and YOLOv8 models. The results show that stain normalization generally improves detection performance in many cases; however, no single approach consistently outperforms others, and its effectiveness depends on the dataset, model, and normalization method.