Clinical Classification-Oriented Ejection Fraction Estimation in Echocardiography Ekokardiyografide Klinik Siniflandirma Odakli Ejeksiyon Fraksiyonu Tahmini
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/siu71813.2026.11636762
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: clinical classification, deep learning, Echocardiography, ejection fraction, left ventricle segmentation, Simpson's disk method
- Orta Doğu Teknik Üniversitesi Adresli: Evet
Özet
Left ventricular ejection fraction (LVEF) is the primary clinical parameter for heart failure diagnosis. Existing automatic LVEF estimation methods generally focus on regression accuracy while neglecting classification performance at clinical thresholds. This study presents a pipeline that estimates LVEF from end-diastolic and end-systolic echocardiography frames with an emphasis on clinical classification performance. The proposed approach employs long axis detection based on mask moments, extended geometric features, and a weighted loss function to address class imbalance. The system is evaluated with two segmentation and three EF estimation methods. Results reveal that the lowest regression error does not guarantee the best clinical classification performance.