1. INTRODUCTION

Image processing has become a popular and critical technology and field of study for our everyday lives. The need to extract important data from visual information arose in many fields like biomedicine, military, economics, industry, and entertainment [1]. Analysis and processing of images requires representing our 3D world in 2D spaces, using different complex algorithms to highlight and examine essential features [2]. With the rapid growth of the volume of visual information, these operations are requiring more computing power. According to Moore’s law, computing performance of classical computers doubles every 18 months. However, experts claim that this law will not hold true for very long [1]. Thus, classical computers will not be able to solve image processing problems with big sets of data within reasonable time limits.

Failure of the Moore’s law can be solved with quantum computation. Existence of more efficient quantum algorithms and their ability to perform calculations faster than classical computers was shown by researchers [1]. Quantum computing also can dramatically improve areas of image processing [2]. Applying quantum computation to image processing tasks is rereferred as Quantum Image Processing or QIMP. This paper will review the basics of quantum image processing and computation, go into its use, with focus on security technologies, and discuss the challenges and future of QIMP.

2. QUANTUM COMPUTATION

A new method of computation known as quantum computing could completely change the field of computer science. In 1982, the late Nobel Prize-winning physicist Richard Feynman began exploring the possibilities of using quantum systems for computing [1]. He was interested in modeling quantum systems on computers. He realized that the number of particles has an exponential effect on the amount of classical memory needed for a quantum system. Thus, when simulating 20 quantum particles, only 1 million values need to be stored, while when simulating 40 quantum particles, 1 trillion values need to be stored. It's impossible to do interesting simulations with 100 or 1000 particles, even with all the computers on Earth [2]. Thus, the concept of using quantum mechanical effects to perform calculations was developed when he proposed the creation of computers that used quantum particles as a computational resource that could model general quantum systems for mass simulation. Researchers have taken a closer look at the processing capability of quantum systems as a result of exponential storage capacity and some disturbing phenomena such as quantum entanglement [4]. Over the past 20 years, quantum computing has exploded, proving that it can solve some problems exponentially faster than any computer [3]. If quantum computers can be built massive enough, the best-known algorithm, Peter Shor's integer decomposition algorithm, will make it easier to break the most common encryption methods currently in use [1].

All modern mainstream computers fall under the category of classical computers, which operate on a "Von Neumann architecture," which is based on an abstraction of discrete chunks of information [1]. Since a computer must eventually be a physical device, scientists recently have moved away from this abstraction of computation and realized that the laws regulating computation should be derived from physical law. One of the most fundamental physical theories, quantum mechanics was a good candidate to investigate the physical feasibility of computational operations [5]. The important finding of this study is that quantum mechanics permits machines that are substantially more powerful than the Von Neumann abstraction.

Along with Shor's factoring algorithm, Lov Grover's search algorithm is a fantastic quantum technique that significantly lessens the amount of work required to look for a certain item. For instance, it takes an average of 500,000 operations on a classical computer to search through a million unsorted names for a given name, and the Von Neumann model of computing offers no faster method [1]. However, using Grover's approach, which takes use of quantum mechanics' parallelism, the name may be obtained with just 1,000 comparisons under the quantum model. Grover's approach outperforms the conventional one considerably more for longer lists.

The subject of quantum computing is huge and diverse today. There are researchers working on a variety of topics, from the creation of physical devices employing various technologies like trapped ions and quantum dots to those tackling challenging algorithmic problems and attempting to pinpoint the precise limits of quantum processing [5]. It has been established that quantum computers are inherently more powerful than classical ones, although it is still unclear how much more powerful they are. And a technological challenge is how to construct a large quantum computer [3].

So, quantum computation is still in its infancy. If the technical challenges are overcome, perhaps quantum computation will one day supersede all current computation techniques with a superior form of computation, just as decades of work have refined the classical computer from the bulky, slow vacuum-tube dinosaurs of the 1940s to the sleek, minimalist, fast transistorized computers that are now widely used. All of this is based on the peculiar laws and procedures of quantum physics, which are themselves anchored in the peculiarities of Nature. What computers will be derived from more complex physical theories like quantum field theory or superstring theory remains to be seen.

3. BACKGROUND

The field of quantum image processing aims to adapt traditional image processing techniques to the quantum computing environment. Its main focus is on using quantum computing technologies to record, modify, and recover quantum pictures in various formats and for various goals. It is believed that QIMP technologies would offer capabilities and performances that are yet unmatched by their traditional equivalents because of some of the astonishing aspects of quantum processing, including entanglement and parallelism. These enhancements could be in the form of increased computer speed, ensured security, reduced storage needs, etc [3].

The first published work connecting quantum mechanics to image processing was Vlasov's work from 1997. It concentrated on using a quantum system to distinguish orthogonal images. Then, efforts were made to look for certain patterns in binary images and identify the target's posture using quantum algorithms. In 2003 publication of Venegas-Andraca and Bose's Qubit Lattice description for quantum pictures greatly contributed to the research that gave rise to what is now known as QIMP. The Real Ket, which Lattorre developed as a follow-up representation, was designed to encode quantum pictures as a foundation for more QIMP applications [1][3].

The proposal of Flexible representation for quantum images by Le et al. genuinely sparked the research in the context of current descriptions of QIMP. This might be explained by the adaptable way in which it enables the integration of the quantum picture into a normalized state, which makes it easier for auxiliary transformations on the image's contents. Since the FRQI, a wide range of computational frameworks that focus on the spatial or chromatic content of the picture have also been presented, along with numerous alternative quantum image representations (QIRs).

The representative QIRs that can be linked back to the FRQI representation include the multi-channel representation for quantum images (MCQI) and novel enhanced quantum image representation (NEQR). The development of algorithms to alter the location and color information encoded using the FRQI and its several variations has also received a lot of attention in QIMP [5]. For instance, it was initially suggested to use FRQI-based fast geometric transformations, which include swapping, flipping, rotations, and restricted geometric transformations to limit these operations to a specific region of an image [3]. Recent discussions have focused on quantum image scaling and NEQR based quantum image translation, which transfer each picture element's position in an input image to a new position in an output image. While single qubit gates like the X, Z, and H gates were initially used to propose FRQI-based broad forms of color transformations. Later, MCQI-based channel of interest operator, which involves moving the preselected color channel's grayscale value, and channel swapping operator, which involves switching the grayscale values of two channels, were further studied [3].

Researchers always prefer to mimic the digital image processing jobs based on the QIRs that we already have in order to demonstrate the viability and competence of QIP methods and applications. Researchers have so far made contributions to quantum image feature extraction, quantum image segmentation, quantum image morphology, and quantum image comparison using the fundamental quantum gates and mentioned operations [5]. QIMP based security technologies in particular have drawn a lot of interest from researchers.

IV. SECURITY TECHNOLOGIES

The necessity for secure communication has developed along with mankind's need to transfer information. With the development of digital technology, the demand for secure communication has increased. In order to realize secure, effective, and cutting-edge technologies for cryptography and information concealment, QIMP is totally based on the extension of digital image processing to the quantum computing domain [3]. Indeed, quantum computation and QIMP offer the potential for secure communication in fields like encryption, steganography, and watermarking.

Encryption is the practice of hiding information to render it unintelligible to those lacking specialized knowledges as a direct application of the science of cryptography. This is frequently done for confidential communications in order to maintain confidentiality. Information hiding focuses on hiding the existence of messages, whereas cryptography is concerned with safeguarding the content of messages. Since attackers cannot easily detect information hidden using techniques like steganography and watermarking, it appears to be safer [3]. The high requirements for the quantity of information that can be concealed under the cover image without changes to its perceived imperceptibility are one of its key limitations, though. Even though steganography and watermarking are similar, they have different goals and/or applications as well as different needs for those goals [3]:

1. In watermarking, the carrier image is the obvious content, but the copyright or ownership is concealed and subject to authentication. In the instance of steganography, it aims to safely transmit the secret message by disguising it as an insignificant component of the carrier image without raising any red flags with outside opponents.
2. Information is concealed through watermarking in the form of a stochastic serial number or an image, such a logo. As a result, watermarked photos typically contain some little copyright ownership information. Steganography frequently needs a huge carrying capacity in terms of the carrier picture because its goal is to conceal the presence of the concealed message.
3. When watermarking, the content can be subject to many sorts of infringements, such as cropping, filtering, channel noise, etc., whereas steganography pictures don’t face such issues.

5. FUTURE DIRECTIONS AND CONCLUSIONS

Research is concentrated on what can be accomplished with quantum technologies once increased realization has been achieved, beyond the continuing work toward the physical implementation of quantum computer hardware [3]. One of these is the nexus of quantum computation with image processing, which is known as quantum image processing. Researchers are confronting both enormous potential and problems to create more effective and usable services because it is a relatively new phenomenon.

All the experimental QIP protocol implementations that have taken place so far have been limited to using traditional PCs and MATLAB simulations built on linear algebra using complex vectors as quantum states and unitary matrices as unitary transforms [5]. These provide a fairly constrained implementation of the potential of quantum computation. Therefore, it is crucial to understand the function of quantum computing software needed to implement the various algorithms that we have in order for them to complement the hardware as researchers intensify their efforts to advance and expand QIP technology [3].

REFERENCES

1. Beach, G., Lomont, C., & Cohen, C. (2003, October). Quantum image processing (quip). In 32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings. (pp. 39-44). IEEE.
2. Anand, A., Lyu, M., Baweja, P. S., & Patil, V. (2022). Quantum Image Processing. arXiv preprint arXiv:2203.01831.
3. Yan, F., Iliyasu, A. M., & Le, P. Q. (2017). Quantum image processing: a review of advances in its security technologies. International Journal of Quantum Information, 15(03), 1730001.
4. Cai, Y., Lu, X., & Jiang, N. (2018). A survey on quantum image processing. Chinese Journal of Electronics, 27(4), 718-727.
5. Ruan, Y., Xue, X., & Shen, Y. (2021). Quantum image processing: opportunities and challenges. Mathematical Problems in Engineering, 2021.
6. Peli, T., & Malah, D. (1982). A study of edge detection algorithms. Computer graphics and image processing, 20(1), 1-21