Decoding Distorted Two-dimensional Barcodes Using Combinatorial Optimization

2D codes are widely used as a shortcut from analog media to online information, and it has been used for authentication of boarding passes of airplanes and trains. Especially in recent years, 2D codes have become popular as a means of cashless payment. However, they are subject to distortion when printed on non-rigid materials, such as papers and clothes. Although general 2D code decoders correct uniform distortion such as perspective distortion, it is difficult to correct non-uniform and irregular distortion of the 2D code itself. This paper proposes a method that can decode 2D codes if they are distorted irregularly and non-uniformly. The proposed method formulates a problem of recognizing the relative position of edges constituting 2D code modules as a combinatorial optimization problem, and locates all modules to reconstruct 2D codes by solving the problem. Experimental results showed that the proposed method decodes the irregularly distorted 2D codes without the help of auxiliary patterns embedded in 2D codes.