The Studies For Aesthetic QR Code Algorithms Based On Base Matrix And Gaussian Model

As one of the entrances to mobile Internet,QR code has some features such as simple generation,large storage capacity and stable scanning.It has become the most widely used two-dimensional code.However,the conventional QR code is made of disorganized black and white blocks,far from aesthetic effect.The QR code beautification technology has attracted more and more researchers.These traditional methods change the color or the shape of QR code modules.These QR codes have poors visual effect and low decoding rate and because their generations do not consider the principles of QR code encoding and decoding.Therefore,it is a key issue how to balance the visual effect and decoding stability in QR code beautification.The principles of QR code encoding and decoding are analyzed.The QR code beautification algorithms are studied.The QR code by the proposed algorithm has a good visual effect and high decoding rate.The main work is as follows:(1)The encoding principles of QR code and RS code are studied,and the forward basis matrix for modifying the filling data area of QR code is constructed.According to the feature of Galois field closing XOR,the reverse basis matrix for modifying the ecc area is constructed using Gaussian Jordan elimination.However,the black and white module flip interacts when the forward basis matrix and the reverse basis matrix are used at the same time.To reduce this influence and make the binary QR code closer to the background image,the visual significance quantization of the module is formulated with the image edge detection,the saliency detection and the module distance weight.Common convex optimization methods cannot solve this formulation because of the mathematical operation of RS code defined in Galois field.So the simulated annealing algorithm is applied to solve the formulation.(2)To achieve a better visual effect,the optimized binary QR code are colorrendered according to the background image.Therefore,a algorithm is proposed to evaluate the attribute of the background image and adjust the brightness according the principles of QR code decoding.The sampling points are mainly concentrated in the central area of the module,and human visual system has a low attention on the smooth area of the image and a high attention on the sharp area.Therefore,The Gaussian model is applied in proposed algorithm for fitting this feature and determining the brightness change of the module.The visual effect and decoding rate is balanced in this way.Image entropy reflects the smoothness of an area on the image,we formalize the parameters of the Gaussian model to a function of image entropy.The parameters of the model are determined with the help of experiment.The visual effect and decoding rate of the proposed algorithm are tested with questionnaire and the decoding experiment.The experiment results show that the QR codes by the proposed algorithm are satisfactory in both visual effect and decoding rate.