Akademik Veri Yönetim Sistemi
28th IEEE Signal Processing: Algorithms, Architectures, Arrangements, and Applications, SPA 2025, Poznan, Polonya, 17 - 19 Eylül 2025, ss.209-214, (Tam Metin Bildiri)
Immunofluorescent Antibody (IFA) tests, used to identify specific antigens in tissue or cell samples, have traditionally been performed manually in clinical laboratories, making them vulnerable to human error and variability. This study presents a deep learning-based approach using the YOLO architecture with the image processing techniques to automate the detection and classification of IFA antibody patterns. We propose two alternative variants of YOLOv8, namely YOLOv8l and YOLOv8x, to recognize 14 antibody patterns defined by International Consensus on ANA Patterns (ICAP). The best performing model, YOLOv8x achieved 90% overall accuracy with mAP@0.5:0.95 score of 53%, indicating robustness across different detection thresholds. Our findings highlight YOLOv8's potential to support more consistent, accurate, and efficient IFA test interpretation in clinical settings.