Correction of geometrical distortions in bands of chromatography images

Abstract

This paper presents a methodology for correcting band distortions in Thin-LayerChromatography (TLC) images. After the segmentation of image lanes, theintensity profile of each lane column is spatially aligned with a reference profileusing a modified version of the Correlation Optimized Warping (COW)algorithm. The proposed band correction methodology was assessed using 105profiles of TLC lanes. A set of features for band characterization was extractedfrom each lane profile, before and after band distortion correction, and was usedas input for three distinct one-class classifiers aiming at band identification. In allcases, the best results of band classification were obtained for the set lanes afterband distortion correction.