Table 6.

Experimental observations of model training, validation, and testing evaluated on Dice and IoU scores. Additional runtime analysis including Inference times, Number of learnable parameters and number of floating-point operations (FLOPs) are also computed for the ablation models.

S.No	Experiment	Dice-coefficient (%)			Mean IoU (%)			Inference Time (milliseconds/image)	Number of Parameters (in millions)	Giga FLOPs
		Training	Validation	Testing	Training	Validation	Testing
1	CCAF upsampler CNN	81.68	80.11	77.67	66.52	65.99	62.65	20.47	18.82	7.04
2	CPA Decoder CNN	85.42	83.76	80.12	73.19	71.61	66.70	26..95	22.45	8.55
3	Proposed CNN	88.01	87.54	85.43	77.57	76.73	73.44	38.24	30.51	13.78