Decoding deformed barcode images with scanpath walkthrough

Abstract

Most camera-based barcode reading systems are scanline-based and designed to work with barcodes attached on flat or smooth surfaces. These systems are likely not able to deal with barcodes on surfaces with high level of deformation, thus making it difficult to find a scanline passing through readable regions. In this paper, an approach that addresses the decoding of deformed barcode images is presented. Our method is based on constructing a scangraph that covers over the area of the barcode, and creating a scanpath that walks through the scangraph by avoiding the regions that are less readable. Barcode digit templates are then used to decode the data sampled along the scanpath. Preliminary results show that some challenging deformed barcode images can be correctly decoded with the presented technique.