UKSimAMSS 22 nd International Conference on Modeling Simulation

UKSim-AMSS 22 nd International Conference on Modeling & Simulation UKSim 2020 QR Keys Generation Based On Chaotic Hybrid Encoding and Fourier Transform Shifting for Image Encryption Walaa Jumiawi Haider Mohammed Presented by Walaa Ali H. Jumiawi

Outlines § Introduction § Related Works § Proposed Method § Experimental Results and Analysis § Conclusions and Future Work UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Image encryption has unique characteristics that differs from other data component, represented by pixels or intensity level distribution. • Pixels representation having undergone many encryption techniques, such as data embedding, AES, chaotic transpositions, DES etc. • Images encryption used over internet, and various applications to ensure higher security and data privacy, such as: - Isolated communication systems - Fingerprints authentications. - Biometric signatures. - Medicals applications. - Military records, and much more. UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Encoding and Entropy: Each Char 16 bit. The number of k ways to be selected from the number of total characters n. • Frequency Domain in Image processing : - Fourier Transform (DFT) • Fourier Transform Shifting by Factor m: - Circular Shift (Circshift) UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Related Works Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • 3 D Chaos Generation Logistic Map: - The parameter μ is taken from range of values. - Initial value needed (range of values) • 3 D Chaos Generation 3 D- Henon chaotic map 3 D-CHEN’s Map: - The parameters x, y and z is taken from range of values. - Initial value needed (range of values) UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Proposed Method Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Keys Encoding and Vector Generation: - Each char encoded dynamically, 16 -bit - No initial values!. - Higher probability of entropy. • Vector to Matrix Reshaping: - This is to generate the map size. -Black is even (0) and white is odd (255). • Fourier Transform Shifting: - Chained QR keys Generation thru shifting - Factor m selected in each chain item. - Mean filter to increase odd even values UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Proposed Method Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Histogram Equalization For Keys: - Equalizing intensities of the filtered QR keys B is any large value to perform equalization • Conditional Shifting Based on QR pixels: - Change pixels location • XORing Operations - Change pixels values UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Experimental Details Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Performance Measures: - Information Entropy - If S is 2^8 char with similar probability, i. e. Si = s 1 to s 256, then the Max entropy of of encrypted image should be closed to P(S) = 8 - NPCR (Number of Pixel Change Rate) - The percentage of different number of pixels in encrypted images, or the pixels change rate. - UACI (Unified Average Changing Intensity) - An average difference of intensity levels in encrypted image - Correlation Coefficient Analysis - The measurement of the quality of pixels pair relocation UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work Experimental Details • Testing three types of images: - Drone Image - To consider the higher level of black intensity level. - Lena Image - To consider the equal gray colors intensity level. - To compare our result with the previous and related methods - Fingerprint Image - To consider the higher level of white intensity level from background Statistical analysis of information entropy, NPCR, UACI, and correlation coefficient. . UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work Experimental Details Methods MMC [14] CCML [13] NPWLCM [23] 3 DCM [5] Proposed Method NPCR UACI Entropy Correlation Coefficient columns rows 99. 5100 33. 4500 7. 9997 0. 0030 0. 0047 25. 0000 19. 0000 Not used 0. 0014 0. 0036 99. 6292 28. 5050 7. 9975 -0. 0195 99. 6048 33. 5044 7. 9890 0. 0014 -0. 0043 99. 62310 33. 95970 7. 99522 0. 00135 - 0. 00422 Comparison using Lena image between previous work and our proposed method. UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Experimental Details Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Correlation Coefficient Analysis - The measurement of the quality of pixels pair relocation in it represents the encryption quality - The left side show the original row/col correlation, and the right side show the encrypted row/col correlation Rows and columns correlations of original and encrypted image UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Conclusion And Future Work Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Conclusion: • Dynamic Encoding can avoid the vulnerability of initial values. • Fourier transform can be used for circular shifting and has been applied in every QR generation step. • Chained QRs key act as a sequence of security layers thru encryption process. • Pixels value coordinates relocation and pixels value transposition produced higher result according to the analysis. • Performance measures showed robustness of the proposed method UKSim-AMSS 22 nd International Conference on Modeling & Simulation

Conclusion And Future Work Introduction Related Works Proposed Method Experimental Results and Analysis Conclusions and Future Work • Future Work: - Mixing previous methods with our proposed method using Fourier transform may produced another hybrid method but the process will be complex. - Testing the proposed method using Crypto-System using machine learning, with big dataset produced from the probability measure in our method. UKSim-AMSS 22 nd International Conference on Modeling & Simulation

UKSim-AMSS 22 nd International Conference on Modeling & Simulation UKSim 2020 Thank You

UKSim-AMSS 22 nd International Conference on Modeling & Simulation UKSim 2020 Thank You