Source-linked AI summary
Cryptanalysis of a Chaotic Image Encryption Algorithm Based on Information Entropy
Chengqing Li, Dongdong Lin, Bingbing Feng, Jinhu Lü, Feng Hao
TL;DR
The paper examines whether IEAIE provides secure image encryption and whether its quantitative security metrics are valid. Through cryptanalysis of the algorithm, it finds defects in its chaotic system, entropy-based sensitivity mechanism, and security evaluation. The authors conclude that IEAIE illustrates common pitfalls rather than a credible secure communication method.
Problem
IEAIE’s claimed security properties and quantitative metrics require scrutiny because image-encryption security can be undermined by defects in chaotic systems and evaluation methods.
Method
The paper cryptanalyzes IEAIE’s structure, digital chaotic map, entropy-based mechanism, and security metrics.
Results
IEAIE’s claimed superiorities are incorrect, and its used security metrics cannot testify to real security performance.
Takeaways & Limitations
IEAIE should be treated as a counterexample illustrating common pitfalls in designing secure multimedia encryption schemes.
Abstract
from arXiv · showhide
Recently, a chaotic image encryption algorithm based on information entropy (IEAIE) was proposed. This paper scrutinizes the security properties of the algorithm and evaluates the validity of the used quantifiable security metrics. When the round number is only one, the equivalent secret key of every basic operation of IEAIE can be recovered with a differential attack separately. Some common insecurity problems in the field of chaotic image encryption are found in IEAIE, e.g. the short orbits of the digital chaotic system and the invalid sensitivity mechanism built on information entropy of the plain image. Even worse, each security metric is questionable, which undermines the security credibility of IEAIE. Hence, IEAIE can only serve as a counterexample for illustrating common pitfalls in designing secure communication method for image data.
I. INTRODUCTION
The paper situates chaos-based image encryption within growing multimedia privacy concerns and introduces its cryptanalytic scrutiny of IEAIE. It targets both the algorithm’s security defects and the validity of its evaluation metrics.
- Growing multimedia collection and transmission increase concerns about privacy once files leave their original control scope.
- Chaos-based cryptography exploits sensitivity and unpredictability for applications including secure communications and pseudorandom number generation.
- Digital degradation of chaotic systems can create security defects that differ from those in non-chaotic encryption schemes.
- IEAIE uses a 2D-LASM-generated pseudorandom sequence to control position permutation and modulo addition, while plain-image entropy affects ciphertext sensitivity.
- For one-round IEAIE, differential divide-and-conquer attacks can separately recover the equivalent secret key of each basic operation.
II. CONCISE DESCRIPTION OF IEAIE
IEAIE represents an image as an M × N 8-bit matrix and encrypts it through chaotic-sequence generation, permutations, gray-distribution changes, and diffusion. The procedure uses two 2D-LASM initial-condition sets as secret-key material and repeats the four main steps for one additional round.
- IEAIE treats image data as an M × N 8-bit matrix rather than using a special image-storage format.
- Key and round structure: The secret key consists of two sets of 2D-LASM initial conditions, and the four encryption steps are repeated for one more round.Decryption reverses the step order and replaces each operation with its inverse.
- Keystream generation: The keystream procedure iterates 2D-LASM for 200 + M·N 2 times, then fills an M × N matrix P from the sequence in raster order.The system uses the 2D-LASM recurrence and a parameter µ drawn from specified intervals.
- Permutation mapping: The transform in Equation (5) cannot always generate a bijective permutation mapping, so it requires correction to ensure successful decryption.The mapping uses u and v values derived from rows and columns of P.
- Encryption procedure: Two vectors derived from chaotic sequences control horizontal and vertical permutations, followed by gray-distribution modification with T and diffusion modulo 256.Horizontal permutation moves columns according to u, vertical permutation moves rows according to v, and diffusion uses K values.
III. CRYPTANALYSIS
The paper challenges the security conclusions previously reported for IEAIE. It attempts to show that the arguments supporting its superior security performance are groundless.
- The paper examines prior analyses that concluded IEAIE has superior security performance.
- The authors attempt to demonstrate that all arguments supporting those security conclusions are groundless.
- The section frames IEAIE’s reported security performance as requiring critical reexamination.
A. SOME SECURITY DEFECTS OF IEAIE
The paper identifies weaknesses in IEAIE’s chaotic map, key design, plaintext-sensitivity mechanism, and security metrics. These issues undermine claims about security, efficiency, and resistance to attacks.
- Underlying chaotic map: 2D-LASM’s digital functional graph forces every orbit into a cycle, while short-cycle regions may leave too few states for the algorithm’s discard strategy.The proposed remedy is an adaptive threshold, but it would add computation.
- Key space and sensitivity: The key-sensitivity test is unconvincing because a small decimal perturbation can change many binary fraction bits, so decimal magnitude does not measure digital initial-condition change.The paper’s binary32 example reports 11 changed fraction bits for the illustrated subtraction.
- Key space and sensitivity: IEAIE’s key-space analysis is unreliable because decimal precision is unsuitable for digital representation and nonchaotic initial-condition regions can yield weak keys.Floating-point representations require explicit mantissa and exponent choices; fixed-point implementations can also make (x0, y0) and (1 − x0, 1 − y0) equivalent.
- Security metrics: Histogram uniformity and variance do not establish security: images can share histograms while differing spatially, and identical zero variances can accompany different visual security.The paper argues that histogram matching should also consider the secret key or plaintext and the corresponding cipher-image histogram.
- Security metrics: Information entropy is unchanged by position permutations and certain value bijections, so many distinct images can share the same entropy and the proposed sensitivity mechanism lacks support.The paper further states that entropy differences in the cited results are bounded by 0.01 and that another permutation-only scheme reaches 7.978.
- Security metrics: Plaintext sensitivity and neighboring-pixel correlation metrics are also questionable because IEAIE’s bit operations propagate changes only from lower- to higher-significance planes and the correlation test lacks a clear decision criterion.The paper additionally notes that deliberately insecure schemes can perform well on the correlation metric.
B. DIFFERENTIAL CRYPTANALYSIS
The paper shows that, for one-round IEAIE, differential attacks can recover the permutation and diffusion equivalents and later decrypt ciphertexts under the same key. The proposed entropy-based keystream dependence does not reliably prevent these attacks because different inputs can produce identical intermediate values or keystreams.
- Attack model: Differential cryptanalysis studies how selected plaintext differences affect corresponding ciphertext differences, enabling analysis of IEAIE under chosen-plaintext access.The attack compares plaintext and ciphertext pairs using invertible difference operations such as bitwise OR or modulo subtraction.
- Entropy-based defense: Different plaintexts can produce the same keystream because entropy computation, parameter mapping, and bit extraction collapse distinct inputs to identical values.The cited causes include modulo operations, quantization error in log2(·), and Eq. (11) extracting only ⌈log2(N)⌉ bits.
- Entropy-based defense: Once plaintext dependence of the keystream is concealed, IEAIE has the same relevant structure as the scheme previously analyzed in, so similar differential cryptanalysis applies.The analysis assumes two plaintexts share the same s and {dj}N.
- Attack procedure: Repeating chosen-plaintext experiments recovers the permutation matrix P and diffusion matrix D as equivalent secret-key components.The procedure isolates one non-zero plaintext difference, locates its permuted position, and then recovers the remaining matrices through the stated equations.
- Attack procedure: The first differing ciphertext elements reveal the permuted location of a single non-zero plaintext difference, as demonstrated on an 8 × 8 example.The two toy plaintexts generate the same s and {dj}8 due to quantization, allowing the location to be observed directly.
- Scope: The paper omits full two-round cryptanalysis because the reported defects are independent of round number and are sufficient to rule out simple fixes.The authors also cite the complexity of presenting the two-round deduction.
IV. CONCLUSION
The paper concludes that IEAIE’s claimed structural advantages and security metrics do not establish real security. It presents the algorithm as a counterexample and calls for broader consideration of multimedia constraints and further cryptanalysis.
- Conclusion: IEAIE’s claimed structural superiorities are incorrect, and every security metric used is incapable of testifying to real security performance.The conclusion states that the algorithm’s security analysis and its quantitative metrics fail to support the claimed security.
- Conclusion: IEAIE can serve only as a counterexample illustrating common pitfalls in designing secure communication methods for image data.The paper’s conclusion frames the algorithm as an example of misleading security evaluation rather than a secure construction.
- Design implications: Secure multimedia encryption design should consider multimedia properties, application constraints, and computation load comprehensively.The conclusion identifies these factors as critical considerations for secure and efficient schemes.
- Design implications: Further cryptanalysis is needed to bridge nonlinear dynamics and modern cryptography.This is stated as a concluding research direction following the analysis of IEAIE.