Structure-aware QR Code abstraction

Quick Response Codes (QR Codes) have proved to be a great success in both automotive industry and general commercial use. However, in most cases, QR Codes are not identified as aesthetical, since the original motivation to invent QR Codes is to achieve high speed and high readability in scanning. This paper proposes a solution framework for QR Code abstraction, which produces machine readable QR Code that is visually similar to the input image. Unlike the existing schemes for QR Code abstraction, which produces QR Codes of low resolution, the framework aims to resolve the high computation complexity of producing halftone QR Codes of high resolution with both similarity and readability preserved. The solution framework has been implemented and tested on various input texts and images, and a user study was conducted to evaluate its performance in preserving similarity. The experimental results show that the framework can produce QR Codes of high resolution and high similarity without compromising readability.