Improving the aesthetic appearance of QR codes

Abstract

This work introduces a method to improve the aesthetic appearance of QR codes subject to bounded probability of detection error. The resultant QR code image embedding is compatible with standard decoding applications for mobile devices available in the market. This method takes advantage of the immunity of QR readers against local luminance disturbances and intends to render the colors of a logo in such a way that the luminance in the dark regions of the codes is detected as dark and the light regions as light. This approach exploits the rich theory and practice of digital halftoning for pixel selection to optimally fuse images into QR codes. The technique minimizes the dual criteria of visibility and decodability. A tractable model for the probability of error is developed. The robustness of the method is adjustable by changing a set of parameters in order to handle variations in the tone given by poor illumination or other factors. The error correction capacity is given by the correction capacity of the QR code. Experimental results are presented evidencing considerable visual quality improvement.