Abstract
QR Code is one of the most popular barcodes and is ubiquitous in our daily life. Being characterized by the massive storage capacity and the ability to be scanned instantly by machine, QR Code is nowadays widely adopted marketing and advertising. The growing demand of QR Codes has driven designer and artists to create a specialized QR codes which incorporate various visual components, e.g. Text, image, or logo, in order to instantly attract people’s attention. However, the inherent nature of QR codes renders the image as seemly noisy black and white squares largely hinders creation of such kind of artistic style QR codes. To resolve the issue, we propose a novel algorithm to combine arbitrary color images with QR codes. The proposed algorithm was designed in a manner that the resultant image not only preserves as much as possible the important content of original image but also keeps the readability of original QR code. The key insight is a representation model inspired by the previous of Chu et al. [2013] which subdivide the QR code module into sub-modules. Based the this model, we propose a revised energy function that works in CIELab color space and takes into account the preservation of module data. The final result is generated by adopting a graph cut algorithm to minimized the energy function. We show the flexibility of the proposed algorithm using several visual pleasing QR codes. We also conducted a user study to validate the readability of resultant QR codes.