Comparison of the results of the fiducial visual source extraction experiment described in Sect. 3 with automatic algorithms. Completeness is shown as a function of S/N_int. The grey points and blue line (here omitting the error bar for clarity; see Fig. 5) shows the fiducial visual experiment, while SoFiA is shown with red lines and GLADoS by green lines. Thick lines show the results after optimising the algorithms to maximise completeness whereas the thin lines are attempts to maximise reliability.