Robust feature extractors based on morphological residues of filtering by reconstruction

We perform a set of texture feature detectors based on morphological residues of opening or closing by reconstruction. This set of feature detectors is proven much more robust to noise than those which based on traditional morphological residues. The robustness of the set of features to noise is investigated in detail qualitatively and quantitatively. It is found that this feature set bears high robustness both to different noise types and noise levels.