QR Code Optimization With Salient Facial Feature

Quick Response Code(QR code) was first invented by Denso Wave corporation and was used for tracking auto parts. It has high reliability, high density of information, fast recognition speed, good anti-jamming, less occupied space and other advantages. Initially, it was mainly used in the product tracking, object recognition, document management and other aspects. In recent years, with the rapid development of mobile technique, QR code has rapidly become the way of information communication for people of all ages, benefit from its simple in form, flexible of 0carrier, advantages of fast transmission of information. However, the traditional QR code is generated without considering visual information. In human eyes, QR code seems to be just black and white "noise" patterns. Therefore, the generating method of visual-pleasant QR code code has become a hot spot of the study of QR code.At present, there are many correlative work in this field. These methods can be divided into 2 kinds: modifying the color and brightness of the original patterns, or replacing the original module of the QR code. The method of modifying the color and brightness of the original pattern can produce visual-pleasant QR code pictures. However, affected by environment light, the decoding efficiency is not guaranteed; the method of replacing original module of QR code is better in decoding efficiedcy, but because of the limitation forms of module, its visual effect is not satisfaction.In order to get visual-pleasant QR code without affecting its decoding efficiency and accuracy, this paper presents a method to generate visual optimized QR Code image with salient facial features, based on module replacement method. The in-puts of this method include a facial image and its corresponding text. Firstly, we generate standard QR Code using the given text. Secondly, we use an iterative FDoG to extract salient facial features. Lastly, we adopt an opti-mization-based pattern replacement algorithm to compute new modules, which are used to replace original ones in QR code. After that, a new QR code image with salient facial features can be generated with these new modules. The QR code generated by this method only using black and white, minimize the influence of ambient light factors, thus it has high decoding efficiency and accuracy; with orienting of significant facial feature, this method can effectively present visual information of original pictures. The experiments show that our method can generate more visual-pleasant QR code images without affecting its decoding rate and accuracy. At the mean time, in order to accelerate the generation speed of QR code, a new method for the optimization of video stream is proposed. Experimental results show that this method can effectively improve the generation speed of the video stream of QR code.