Conclusion and Future Work
Creating visually attractive QR codes has received a growing demand in diverse fields especially in marketing and advertising.
The main challenge lies in how to incorporate high-level visual features
(e.g. colors, letters, illustrations, logos, etc.) with ordinary QR code without compromising its readability. We presented an automatic algorithm to create a new type of visual QR code, called halftone QR code, at a controllable level of readability. There are two major components in our approach. One is a representation model that
minimally binds to the original QR codes and is flexible to adapt
to target halftone images; and the other is an introduction of pattern reliability which enables the control over level of readability.
We proposed a systematic approach to automatically build a large database of synthetic QR codes and used it to statistically model the pattern reliability. Finally, we tested our approach using a wide variety of images and conducted an experiment of manually scanning the resultant QR codes to validate the effectiveness of our approach.
In addition to generating valid halftone QR codes, we also believe that our synthetic database is a potential benchmark tool for evaluating the performance of various barcode readers.