The measured sample will sometimes contain areas that do not belong to the actual sample. In these cases, it may be valuable to remove the non-sample areas from further modeling. The first step will be to create a PCA model of the imported image data. A selection corresponding to the non-sample areas can be then decided by trying different selection in a Scatter 2D density plot of the scores and at the same time inspect the interaction to the RGB image. These two plots can be created by dragging the scores, T, from the Data Tree and then choose the appropriate items from the list. When the correct selection is made, the entire non-sample area will be selected in the RGB image.

The selected pixels can be excluded from the model by right-clicking on the Scatter 2D plot and then choose "Exclude" from the plot menu. The model will be recalculated when "Apply Changes" is chosen from the plot menu. The plots will be updated automatically and only the pixels belonging to the actual sample will remain.