Table 4.

Observations of the Ablation studies. The results are grouped dataset-wise. The proposed models were trained and tested under each dataset.

S No	Dataset	Method	DSC	IoU	Precision	Sensitivity	Specificity	AUC
1	Jun Ma dataset [30]	CCAF upsampler CNN	80.43	69.87	78.85	82.46	99.75	77.00
		CPA decoder CNN	82.63	72.20	80.26	96.06	99.76	81.69
		Proposed CNN	88.01	75.03	85.57	90.05	99.77	85.03
2	MosMedData [31]	CCAF upsampler CNN	75.19	65.30	73.11	77.32	99.70	77.34
		CPA decoder CNN	83.49	72.78	84.66	82.19	99.75	83.26
		Proposed CNN	83.71	71.51	82.43	84.58	99.75	82.21
3	Combined dataset	CCAF upsampler CNN	77.67	62.65	76.32	79.21	99.72	75.60
		CPA decoder CNN	80.12	68.70	80.96	79.39	99.79	78.69
		Proposed CNN	85.43	73.44	81.23	89.88	99.77	84.61