Table 4.

Comparison of the severity quantification performances of single-task model for classification and multi-task model for two tasks.

Models	External dataset 1(CNUH)	External dataset 2(YNU)	External dataset 3(KNUH)
	MSE (95% CI)	MSE (95% CI)	MSE (95% CI)
Multi-task model	1.441(0.760-2.122)	1.435(1.195–1.676)	1.458(1.147–1.768)
Single-task model	1.645(0.969–2.320)	1.417(1.138–1.695)	$1.{731}^{**}$(1.372–2.090)

Note:${}^{**}$ denotes the better performance of our model with statistical significance ($p<0.001$). CI, confidence interval.