Mean-median correlation is computed for each spot in an array by
dividing the smaller of the mean or median of the signal intensities
by the larger. High quality spots should show a high value, according
to [3]. Hence you can give less weight to spots with low
values. The mean-median correlation is easy to compute, but in order
to detect outliers, it must be combined with a saturation
filter. Doughnut shapes are also difficult to detect using mean-median
correlation.
Bjørn Kåre Alsberg
2006-04-06