Table 6.

Results of the K-Fold cross-validation for the baseline on Task 1 training dataset (as reported in our earlier data analysis paper (Bano et al., 2021), included here for completeness) Mean IoU for each class over each video and, in the last row, the average mean IoU per class are reported. Key: BG-background.

Video	Class				Overall	Fold	Images	Class				Overall
	BG	Vessel	Tool	Fetus	per video		per fold	BG	Vessel	Tool	Fetus	per fold
Video001	0.83	0.85	0.69	0.74	0.64	1	352	0.80	0.83	0.64	0.74	0.61
Video006	0.67	0.67	0.74	0.76	0.58
Video016	0.80	0.83	0.64	0.74	0.60
Video002	0.78	0.79	0.80	0.53	0.56	2	353	0.80	0.81	0.83	0.78	0.69
Video011	0.75	0.72	0.73	0.83	0.64
Video018	0.80	0.81	0.83	0.78	0.71
Video004	0.80	0.80	0.72	0.80	0.66	3	349	0.76	0.78	0.79	0.55	0.65
Video019	0.81	0.81	0.64	0.85	0.65
Video023	0.76	0.78	0.79	0.55	0.56
Video003	0.79	0.81	0.72	0.79	0.66	4	327	0.82	0.82	0.80	0.93	0.66
Video005	0.71	0.77	0.79	0.56	0.56
Video014	0.82	0.82	0.80	0.93	0.78
Video007	0.78	0.77	0.84	0.72	0.66	5	350	0.78	0.81	0.85	0.54	0.67
Video008	0.78	0.76	0.75	0.85	0.68
Video022	0.78	0.81	0.85	0.54	0.60
Video009	0.80	0.80	0.80	0.73	0.66	6	329	0.66	0.66	0.73	0.57	0.58
Video013	0.72	0.77	0.75	0.50	0.50
Video017	0.66	0.66	0.73	0.57	0.48
per class	0.78	0.79	0.76	0.75