Abstract: The existing aesthetic QR code algorithms do not consider the region of interest of a background image, which affects the beautification effect. A new aesthetic QR code algorithm based on region of interest and RS code mechanism was proposed. Firstly, an improved region of interest detection algorithm based on multiple features was proposed to get the salient binary image of the background image. Secondly, an intermediate QR code was obtained by performing an XOR operation on the original RS code by using the RS coding matrix, which was completely consistent with the salient binary image of the background image. Then the background image and the intermediate QR code were fused according to a specific fusion method. To further expand the aesthetic area, the RS error correction mechanism was performed on the fusion map, getting the final aesthetic QR code image. Experimental results on the established test sample set show that the proposed aesthetic algorithm can achieve complete background replacement, save more image information, and has a better visual effect and a higher decoding rate.

Key words: Region Of Interest (ROI), multi-feature, Reed-Solomon code, RS coding matrix, aesthetic QR code

摘要： 针对已有QR码美化方法没有考虑背景图像的感兴趣区域，进而影响美化效果的问题，提出基于感兴趣区域和RS编码机制的QR码美化算法。首先提出改进的基于多特征感兴趣区域检测算法，进而用此方法得到背景图的显著二值图。其次，将原始QR码利用RS编码矩阵进行异或操作得到中间QR码，该QR码和背景图的显著二值图完全一致；然后，将背景图像和中间QR码按照特定的融合策略进行融合。最后，将融合图再次利用RS纠错机制进一步扩大美化区域，得到最终的美化QR码图像。在测试样本集上的实验表明:所提算法可以实现完整的背景替换，保存更多的图像信息，具有较好的视觉效果和较高的解码率。

关键词: 感兴趣区域, 多特征, Reed-Solomon纠错编码, RS编码矩阵, 美化QR码