Identification of Antibody Patterns in Immunofluorescence Antibody Tests Using a YOLO26-Based Architecture YOLO26 Tabanli Mimari ile Immünofloresan Antikor Testlerinde Antikor Desenlerinin Tanimlanmasi
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/siu71813.2026.11636568
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: Artificial Intelligence, Autoimmun disorder, Deep Learning, IFA test, YOLO26
- Trakya Üniversitesi Adresli: Evet
Özet
Immunofluorescence Antibody (IFA) tests are methods used to detect specific antigens in tissue or cell samples and are traditionally performed manually in clinical laboratories. This makes it susceptible to human error and observer-dependent variability. This study presents a deep learning approach based on the YOLO architecture, supported by image processing techniques, for the automated detection and classification of IFA antibody patterns. Within the proposed method, 15 different antibody patterns were classified in the dataset, and the results were obtained according to the YOLO26 architecture with an F1 score of 0.887 and an average precision of (mAP50) 0.937. These results demonstrate that the model exhibits stable performance at different detection thresholds.