A Study on Digital Image Authentication and Recovery Using 2D QR Codes

• Main Authors: LIN, Tzu-Wei, 林子瑋
• Other Authors: Wu, Wen-Chuan
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/56871403842520088790

碩士 === 真理大學 === 資訊工程學系碩士班 === 103 === Image authentication is an important technology to protect image content integrity. Most of these methods use self-embedding technique, which embeds particular authentication data into the image itself. The embedded result is named an authenticated image, which must have very good quality such that it cannot perceive any differences between two highly similar images with the naked eye. When that image was illegally modified during the transmission via the public network, an image authentication method is able to extract the hidden data to determine whether an image has been tampered with or not. Further, the content of tampered regions could be restored back to its initial state in order to reduce the probability of image retransmission and save network bandwidth. This thesis proposed three image authentication methods. The first method is to adopt the stable singular values as authentication data and then hide them randomly into image pixels. However, false tampering detection might occur when some areas were tampered in such a way that the authentication data of the others make changes. In order to prevent this defect, therefore, this thesis attempt to represent the authentication data into a quick response (QR) code in the second method. The main reason is that a QR code possesses the capability of strong tolerance and error correction. It allows a QR code to be read correctly even if the code is slightly dirty or damaged. Hence, this property is applied to protect the authentication data and further recover the potentially tampered regions if necessary. Finally, we also applied that code to vector quantization compression to improve the previous methods with good quality and few false positives.