Table 4.

Per Class Average recall for CarLab 2025 and other frameworks with in-domain testing (%).

Label	CarLab 2025	Narrow	da Silva Moura 2024	USVAC 2025 Level 1	USVAC 2025 Level 2
Functional Dysphonia	$51\pm 20$	$7\pm 6$	$47\pm 18$	$39\pm 16$	$28\pm 22$
Hyperfunctional Dysphonia	$54\pm 8$	$37\pm 13$	$54\pm 12$	$28\pm 13$	$36\pm 8$
Laryngitis	$60\pm 14$	$38\pm 10$	$25\pm 15$	$63\pm 18$	$47\pm 17$
Leukoplakia	$58\pm 14$	$6\pm 7$	$44\pm 22$	$71\pm 23$	$16\pm 12$
Normal	$75\pm 16$	$81\pm 12$	$83\pm 7$	$84\pm 9$	$78\pm 10$
Psychogenic Dysphonia	$43\pm 14$	$17\pm 10$	$40\pm 11$	$34\pm 9$	$28\pm 16$
Reinke’s Edema	$45\pm 8$	$25\pm 12$	$43\pm 19$	$63\pm 15$	$27\pm 14$
Recurrent Paralysis	$55\pm 11$	$50\pm 19$	$48\pm 17$	$65\pm 16$	$46\pm 12$
Vocal Fold Polyp	$54\pm 12$	$10\pm 8$	$40\pm 15$	$68\pm 9$	$19\pm 14$
Average	$55\pm 13$	$30\pm 11$	$47\pm 15$	$57\pm 14$	$36\pm 14$

Here narrow classification includes only the disorders listed in this table.