Modulus Calculations on Prime Number Algorithm for Information Hiding With High Comprehensive Performance

Yu Zhang, Sha Wang, Teng Li, Bo Liu, Dongbo PanIEEE Access 2020 · DOI

Abstract

Information hiding is an important technique for information security, which is wildly studied by researchers. Recently embedding methods are proposed in spatial, frequency and other domains. After investigating previous literatures, we find that there is still room for embedding performance improvement. Inspired by some literatures, we propose a new method (Modulus Calculations on Prime Number Algorithm, MOPNA) for embedding secret data into cover-images. The main idea of MOPNA is hiding confidential data in paired cover-pixels with modulus calculation based on weight parameters consisting of prime numbers. MOPNA improves the embedding capacity while maintaining good stego-image quality. The correctness of MOPNA method is proved by a combination of mathematical and programming proof. The experimental results prove that the proposed method has high embedding capacity and achieves better comprehensive performance than existing methods.

Introduction

In the field of information security, steganography serves as a critical method for concealing the existence of secret data within a digital medium, such as an image. The primary challenge in developing effective steganographic algorithms is the inherent trade-off between two competing objectives: capacity and imperceptibility. Capacity refers to the volume of secret data that can be embedded per pixel, while imperceptibility measures how well the "stego-image" maintains the visual and statistical properties of the original cover image.

While frequency-domain methods offer robustness, they often suffer from low capacity. Spatial-domain methods provide higher capacity but frequently struggle to maintain high visual quality (measured by Peak Signal-to-Noise Ratio, or PSNR) as the payload increases. This paper introduces the Modulus Calculations on Prime Number Algorithm (MOPNA) to bridge this gap, aiming for high comprehensive performance across both metrics.

Methodology: The MOPNA Algorithm

The core of MOPNA lies in its use of number theory to optimize data embedding within pixel pairs. The algorithm follows a structured process to minimize visual distortion while maximizing data throughput.

Pixel Pairing and Prime Weighting

The algorithm divides the cover image into pairs of adjacent pixels $(g_0, g_1)$. It utilizes a weight parameter $C = (c_0, c_1)$ consisting of prime numbers—specifically the pair 3 and 11—to ensure a uniform distribution of possible outcomes during the modulus operation. This mathematical property is key to finding a "nearby" pixel state that matches the secret data with minimal modification.

Modulus Calculation and Optimization

Secret data is converted into a decimal value $d$. MOPNA then identifies a small modification $X = (x_0, x_1)$ to the original pixels such that the following modulus function is satisfied:

$$(c_0(g_0+x_0) + c_1(g_1+x_1)) \mod 2^{2k+1} = d$$

To ensure the highest possible image quality, MOPNA employs a novel Changes in Pixel Values (CPV) function. Since multiple modifications might satisfy the modulus equation, the CPV function selects the specific modification that results in the smallest visual change. The authors provide both mathematical theorems and programming-based proofs to guarantee that a valid modification always exists for secret data within the defined ranges.

Experimental Setup

The performance of MOPNA was evaluated using a standard suite of 256×256 grayscale benchmark images. These images, including Baboon, Barbara, Boat, F-16, Goldhill, Lena, Pepper, and Tiffany, are widely recognized in image processing research for their varying levels of texture and detail.

The Baboon benchmark image, a standard test case for evaluating steganographic algorithms.

Results were compared against several established algorithms, such as EMD06, LWC07, JY09, GEMD13, KKWW16, and SB19, to establish a baseline for comprehensive performance.

Key Results and Discussion

High Embedding Capacity

MOPNA demonstrates significant capacity, achieving up to 3.5 bits per pixel (bpp) when the parameter $k=3$. This allows for substantial data payloads without requiring excessively large cover images.

Quality-Capacity Balance

When hiding a fixed payload of 49,000 bits, MOPNA maintained a PSNR of 37.93 dB at its highest capacity setting. While some older methods like EMD06 may show higher PSNR at lower capacities, they lack the efficiency of MOPNA, which requires fewer pixels to hide the same amount of data.

Comprehensive Performance Index (RPN)

The authors propose the RPN (Ratio of PSNR and number of pixels used) as a more holistic metric for steganography. This index rewards algorithms that achieve high visual quality while utilizing fewer pixels. In these tests, MOPNA consistently emerged as the leader. For a 98k bit payload, MOPNA achieved an RPN of 1.35, outperforming its closest competitor, SB19, which scored 1.22.

Conclusion

MOPNA represents a significant advancement in spatial-domain steganography by leveraging prime-number-weighted modulus operations. By treating data embedding as an optimization problem, it achieves a superior balance of capacity and imperceptibility. Future research will focus on testing MOPNA against advanced statistical steganalysis tools and extending the mathematical proofs to support even higher embedding capacities ($k \ge 6$).

Summary updated

Abstract

Information hiding is an important technique for information security, which is wildly studied by researchers. Recently embedding methods are proposed in spatial, frequency and other domains. After investigating previous literatures, we find that there is still room for embedding performance improvement. Inspired by some literatures, we propose a new method (Modulus Calculations on Prime Number Algorithm, MOPNA) for embedding secret data into cover-images. The main idea of MOPNA is hiding confidential data in paired cover-pixels with modulus calculation based on weight parameters consisting of prime numbers. MOPNA improves the embedding capacity while maintaining good stego-image quality. The correctness of MOPNA method is proved by a combination of mathematical and programming proof. The experimental results prove that the proposed method has high embedding capacity and achieves better comprehensive performance than existing methods.
BibTeX
@inproceedings{zhang2020modulus,
  author = {Yu Zhang and Sha Wang and Teng Li and Bo Liu and Dongbo Pan},
  title = {{Modulus Calculations on Prime Number Algorithm for Information Hiding With High Comprehensive Performance}},
  booktitle = {IEEE Access},
  year = {2020},
  doi = {10.1109/access.2020.2992690},
}
Figure