A Novel Technique for Web Page Informative Content Extraction

This paper presents an automated method for detection and removal of cracks in digitized colour paintings. The algorithm starts with the extraction of crack centerlines, which are used as guidelines for the subsequent crack- filling phase. For this phase, the output of four direction differential operators are processed in order to select connected sets of candidate points to be further classified as centerline pixels using crack derived features. The final segmentation is obtained using an iterative region growing method that integrates the contents of several images resulting from crack width dependent morphological filters. The methodology has been shown to perform very well on digitized paintings suffering from cracks.