Fast Level Line Transform



The Fast Level Line Transform (FLLT) constructs a contrast-invariant representation of an image. This algorithm builds a tree which follows the inclusion of the shapes contained in an image. For an image filter, having the contrast-invariant property is interesting. For instance, in the field of document image analysis, this representation is precious to extract characters whatever their mean gray-levels are brighter or darker than their surroundings. This document presents how this algorithm is introduced in our image processing library and shows the results of some connected that can be derived from this representation.