Optics-based system for robust and reliable encryption
The safe storage and transmission of data continues to be a challenge because of increasing threats to the confidentiality, integrity, and availability of information such as personal identification data or biometrics images.
Security systems based on optoelectronics can perform highly accurate encryption and decryption in almost real time. A classical optical system employs two random phase masks in the spatial (space domain) and Fourier planes (frequency domain).1Several other optical systems have also been proposed, such as double phase encoding using joint transform correlation (JTC, where both input and reference are introduced jointly),2 exclusive-OR encryption,3 security verification using a polarization-encoded mask,4 a multiplexed minimum average correlation energy phase-encrypted filter,5 and fractional Fourier transformation.6, 7
However, the majority of these techniques require complex transformation operations, have random phase keys that are vulnerable to being hacked or guessed, or suffer from correlation problems. Encryption and decryption are based on correlation between the image/information and the secret key, and some of these techniques cannot produce sharp correlation signals.
We have developed a novel optical information security system that employs orthogonal encryption codes (which offer no cross-correlation among themselves) within the context of a multiple phase-shifted reference-based JTC (MRJTC) technique.8Our system is an efficient cryptographic tool for biometric security and similar fields and offers robust security and faithful reproduction of the original data on demand.
The proposed optical information security system is shown in the block diagrams of Figures 1 and 2. The given input images containing confidential information are linearly encoded using individual orthogonal codes, and then superimposed on a common spatial domain: see Figure 1(a). The encoded and multiplexed image is next nonlinearly encrypted using another key. First, the encryption key is fed to four parallel processing channels with phase shifting by 0°, 90°, 180°, and 270°. Then the encoded and multiplexed image is introduced to each of the phase-shifted keys to form four joint images. The joint images are then Fourier transformed independently and from their transform the magnitude spectra are recorded as four joint power spectra (JPS) signals. These are phase-modulated and then combined to form a modified JPS signal. An inverse Fourier transform of this yields the encrypted image: see Figure 1(b).
For decryption, the received image is first Fourier transformed and then multiplied with the Fourier transformation of the encryption key. Inverse Fourier transformation of the resulting signal retrieves the original encoded image: see Figure 2(a) The decryption process does not require any complex conjugate of the encryption key and also yields an uncorrupted input image as the output does not contain any unwanted correlation terms. The correlation process involving Fourier transformation generates some noisy signals around the correlation peak signal, called fringes. To further enhance the decrypted image, we apply a ‘fringe-adjusted’ filter to eliminate this noise and make the correlation output sharp and distinct. Then we apply the inverse Fourier transformation.
Finally, the decrypted image is decoded with the corresponding individual orthogonal code: we perform a threshold operation to recover the original image, where the threshold value can be selected based on the format of the orthogonal code set used: see Figure 2(b).
We evaluated the performance of the optical security system by computer simulation using MATLAB software.9 Binary images of 32×32 pixels were employed including alphanumeric characters. We used a Walsh code of length 4 to encode the images.
Figure 3(a) shows the input image to be encrypted, which includes 4 different characters. The encoded and multiplexed image is shown in Figure 3(b), which includes the information from all four input image characters. We then applied the MRJTC techniques to perform nonlinear encryption on this image. From the encrypted image in Figure 3(c) it is obvious that the original information is completely scrambled and hidden in the output image plane. This ensures enhanced security.
Another very important requirement for any security system is to correctly reproduce the information without any loss. Figure 3(d) shows the result of the first decryption process on the received encrypted image, indicating perfect reproduction of the encoded and multiplexed image when compared with Figure 3(b). Finally, the orthogonal code decoding process retrieves the individual characters from the encrypted image without any loss or distortion, as is obvious from Figure 3(e).
We also investigated the security strength of our proposed encryption and decryption technique. We attempted to decode the confidential information without knowing the correct codes. Figure 3(f) shows the result of the MRJTC decryption process using a wrong code. Neither useful information nor the encoded image are obtained. Finally, using the decoding process with another wrong code produces individual character information as shown in Figure 3(g). Obviously, the use of orthogonal codes and the nonlinear MRJTC encryption process have prevented any part of the information from being decoded, and there are no clues as to the information content.
We simulated and tested our technique under various challenging conditions to investigate its robustness against security attacks. In every case, it performed with efficiency and reliability. The technique offers a simple architecture eliminating the requirement for any complex conjugate of the address code to decrypt the input information. Our proposed system fulfils the requirements for confidentiality and information integrity, at the same time reproducing the information for an authorized user with 100% accuracy. The system offers an ideal and practical candidate for a security system to protect the storage and transmission of confidential information.
Old Dominion University
Mohammad Karim is vice-president for research and professor of Electrical and Computer Engineering. He has authored or edited 19 books, 33 journal special issues, over 375 research papers, and eight book chapters. He is a fellow of SPIE, the Optical Society of America, the IEEE, the Institute of Physics, the Institution of Engineering & Technology, and the Bangladesh Academy of Sciences.
Farmingdale State College
State University of New York
Mohammed Nazrul Islam is an assistant professor. Prior to joining Farmingdale, he worked as a research scientist and adjunct assistant professor in the Department of Electrical and Computer Engineering at Old Dominion University, VA. He has also worked as a visiting assistant professor at the University of West Florida and as a postdoctoral research fellow at the University of South Alabama. He received Bachelor of Science and a Master of Science degrees in Electrical and Electronic Engineering from Bangladesh University of Engineering and Technology in 1991 and 1994, respectively. He then earned a PhD from Muroran Institute of Technology, Japan, in 1999. He has authored and coauthored more than 120 publications in refereed journals and conference proceedings. He is a senior member of the IEEE and also of SPIE.
University of Dayton (UD)
Vijayan Asari is a professor in Electrical and Computer Engineering and the Ohio Research Scholars Endowed Chair in Wide Area Surveillance at UD. He is the director of the Computer Vision and Wide Area Surveillance Laboratory (Vision Lab) at UD. He received his Bachelor's degree in electronics and communication engineering from the University of Kerala, India in 1978, and MTech and PhD degrees in electrical engineering from the Indian Institute of Technology, Madras in 1984 and 1994, respectively. He joined the University of Dayton in February 2010. Prior to joining UD, he worked as a research fellow at the National University of Singapore, as a research fellow at Nanyang Technological University, Singapore, and as a professor in Electrical and Computer Engineering at Old Dominion University (ODU), VA. He was the founding director of the Computational Intelligence and Machine Vision Laboratory (ODU Vision Lab) at ODU. He holds two patents and has published more than 280 research papers, including 58 peer-reviewed journal papers. His research areas include image processing, computer vision, machine learning, and pattern recognition. Vijayan Asari is a senior member of the IEEE and SPIE.
