Develpment of data embedding and extracting for printed images considering mobile devices

在本研究中，我们开发了一种利用移动设备中的相机对印刷图像进行数据嵌入和数据提取的方法。我们研究了几何畸变和透镜畸变的校正方法，可以大大减少加工时间。此外，我们还开发了一种从多幅捕获的图像中只选择状态较好的图像的方法，提高了嵌入数据的检测率。通过在平板电脑上的实现，研究了图像采集接口，并用C语言重写了所有程序。实验表明，对于PSNR为40dB、嵌入数据为112位的图像，检测率达到95%。特别是在良好的条件下，检测率达到97%以上。该嵌入方法的改进还可以实现嵌入嵌入数据的192位，对应于48个字符的字母和数字。

In this study, we develop a data embedding method to a printed image and a data extraction method by using a camera in mobile devices. We investigate the correcting method for geometrical and lens distortion and can drastically reduce the processing time. Additionally, we improve the detection rates of embedding data by developing a method which selects only the image in a good condition from several captured images. Through the implementation to a tablet PC, the interface for capturing the images is studied and all programs are rewritten by C language. From the experiments by using this tablet, we obtained the detection rate of 95% for the image which has 40dB of PSNR and 112 bits of the embedding data. Especially, the detection rate of over 97% is obtained under the good condition. The improvement of the embedding method can also achieve to embed the 192 bits of the embedding data which correspond to 48 characters of alphabets and numbers.