Cross talk reduction of the three compared unmixing methods for both the synthetic and real images.

The table shows the p-values obtained when evaluating the average decrease in crosstalk between each pair of unmixing algorithms using a paired one-tailed non-parametric Wilcoxon ranked-sign test. The null-hypothesis in each case is that the change is not significant while the alternative hypothesis is that the algorithm in a given row is significantly worse than the corresponding column algorithm.