論文ID: e26.68
Feline heart disease is a primary cause of sudden mortality in cats, necessitating early detection and routine monitoring of heart murmurs via auscultation. However, because cats are sensitive to environmental changes, clinic-induced stress can trigger behavioral alterations and physiological anomalies, such as heart rate deviations, thereby confounding accurate diagnosis. Although home auscultation offers a low-stress alternative, purring frequently obscures cardiac signals. To address this limitation, this letter proposes a method to isolate heart sounds, including heart murmurs, by removing purring noise from auscultatory signals using convolutive non-negative matrix factorization.