Hiding Information in Image Frequency Domain

Before the Qingming Dynasty, my classmates forward a link to me, it’s a question list. One of the related topics was to give two pictures, the original one and the processed one. They asked to decrypt the information.


The title gave a hint to Fourier, so it is easy to guess that digital watermarking technology which based on FFT. However, I didn’t find this prompt at first, so I made some attempts according to my experience. First, I considered whether it was a special kind of image, that is, hiding the information at the end of the image file by using the characteristics of image coding format and compressed file with it.

After failure, I considered whether it was hiding the bit that did not affect the image display by using the characteristics of PNG.

And, it fail again, I also consider LSB and it still not work.

Finally, I compare the two pictures with differ command, find that the different points are distributed in four corners of the image. Then I compare the specific content of the two PNG format pictures with WinHex. It is found that from the first IDAT, the hexadecimal value and CRC are different, excluding the simple embedded text. Then I found the hint, and then it was easy.

Using MATLAB to do FFT on the image, and then subtraction can get a more noisy two-dimensional code, multiplied by the parameters can get a clear two-dimensional code.


I was not sensitive to the image because did not take the course of digital image processing. When I found the difference, I did not immediately respond that using frequency domain to hide the information. That’s take too many time.

Reference Material

  1. 阿里巴巴公司根据截图查到泄露信息的具体员工的技术是什么? (What is the technology of Alibaba based on the screenshots to find out the specific employees who disclose the information?)
  2. 纪念我曾经遇到的那些脑洞大的要死的隐写术 (To commemorate the death of the brain cave I encountered.)
  3. CTF中图片隐藏文件分离方法总结 (Summary of image hiding file separation method in CTF)
  4. 数字图像加密算法之空域LSB (Spatial domain LSB for digital image encryption algorithm)
  5. 2016 HCTF Crypto 出题总结 (Summary of the problem In 2016 HCTF Crypto)