Fig. 3.

Quantitative performance of the DLE-standard model evaluated on the testing set in terms of lesion SUV_max, lung SUV_mean, liver SUV_mean and noise in liver for full-, ¾-, ½- and ¼-duration input scans. Average of STD is the SD of noise averaged over all 5 liver VOIs across all patients. The p values shown are calculated using the Wilcoxon signed-rank test from the differences in SUV bias (OSEM to DLE-standard). BSREM = block sequential regularised expectation maximisation, DLE = deep learning enhancement