Enhanced security for medical images using a new 5D hyper chaotic map and deep learning based segmentation

Abstract

Medical image encryption is important for maintaining the confidentiality of sensitive medical data and protecting patient privacy. Contemporary healthcare systems store significant patient data in text and graphic form. This research proposes a New 5D hyperchaotic system combined with a customised U-Net architecture. Chaotic maps have become an increasingly popular method for encryption because of their remarkable characteristics, including statistical randomness and sensitivity to initial conditions. The significant region is segmented from the medical images using the U-Net network, and its statistics are utilised as initial conditions to generate the new random sequence. Initially, zig-zag scrambling confuses the pixel position of a medical image and applies further permutation with a new 5D hyperchaotic sequence. Two stages of diffusion are used, such as dynamic DNA flip and dynamic DNA XOR, to enhance the encryption algorithm's security against various attacks. The randomness of the New 5D hyperchaotic system is verified using the NIST SP800-22 statistical test, calculating the Lyapunov exponent and plotting the attractor diagram of the chaotic sequence. The algorithm validates with statistical measures such as PSNR, MSE, NPCR, UACI, entropy, and Chi-square values. Evaluation is performed for test images yields average horizontal, vertical, and diagonal correlation coefficients of -0.0018, -0.0002, and 0.0007, respectively, Shannon entropy of 7.9971, Kolmogorov Entropy value of 2.9469, NPCR of 99.61%, UACI of 33.49%, Chi-square "PASS" at both the 5% (293.2478) and 1% (310.4574) significance levels, key space is 2⁵⁰⁰ and an average encryption time of approximately 2.93 s per 256 × 256 image on a standard desktop CPU. The performance comparisons use various encryption methods and demonstrate that the proposed method ensures secure reliability against various challenges.

Keywords: Critical region segmentation; Dynamic DNA encoding; Hyper-chaotic map; Image encryption; U-Net; Zig-zag scrambling.

Fig. 1

Attractor diagram for NEW 5D Hyper chaotic system.

Fig. 2

2D-time series plot: (a) 2D-time series plot for chaotic sequence-X, (b) 2D-time series plot for chaotic sequence-Y (c) 2D-time series plot for chaotic sequence-Z.

Fig. 3

(a) Bifurcation diagram of New 5D Hyperchaotic system, (b) Lyapunov exponent of 5D chaotic system concerning time (c) Lyapunov exponent of 5D chaotic system to the control parameter α.

Fig. 4

(a) Kolmogorov entropy of new 5D Hyperchaotic system.

Fig. 5

U-NET Architecture.

Fig. 6

Proposed block diagram.

Algorithm 1: Key Generation

Algorithm 2: Function: encrypt_medical_image

Fig. 7

Zig zag scrambling.

Algorithm 3: Zigzag_scramble (image, iterations)

Algorithm 4: Permute_pixels (image, sorted_indices)

Algorithm 5: Encode_image_to_dna (image, dna_rules).

Algorithm 6: Perform_dna_flip(dna_image, flip_positions)

Fig. 8

(a) Training and validation accuracy of the U-NET network for the RITE dataset (b) Training and validation loss for the RITE dataset.

Algorithm7: dna_xor_diffusion(dna_image, chaotic_dna_sequence)

Algorithm 8: Function: decrypt_medical_image

Fig. 9

Drive data set -Test image with predicted retinal Image from U-NET.

Fig. 10

Rite data set -Test image with extracted retinal Image from U-NET.

Fig. 11

(a)original image, (b) key image, (c) zig-zag scrambled image, (d)confused image, (e) encrypted image (f) decrypted image (g) original image histogram, (h) key image histogram, (i) zig-zag scrambled image histogram, (j) confused image histogram, (k) encrypted image histogram, and (l) decrypted image histogram.

Fig. 12

(a) original eye image, (b) predicted retinal image, (c) zig-zag scrambled image, (d) confused image, (e) DNA swapped image (f) chaotic sequence ‘V’-DNA image, (g), DNA XOR image (Final cipher image) and (i) decrypted image.

Fig. 13

(a) Horizontal correlation of original image and encrypted image, (b) vertical correlation of original image and encrypted image, and (c) Diagonal correlation of original image and encrypted image for Medical Drive test image_1.

Fig. 14

Histogram analysis of medical images. (a), (b), (c), (d), (e) are the 5 test medical images and (f), (g), (h),(i),(j) the histogram of the input medical images, (k), (l), (m), (n), (o) are the histogram of encrypted medical images.

Fig. 15

Cropping Images (a) and (b) 5% and 25% vertical cropping image, (c) and (d) 5% and 25% Horizontal cropping image, (e) and (f) 5% and 25% center cropping. (g) and (h) decrypted image of (a) and (b), (i) and (j) decrypted image of (c) and (d), and (k) and (l) decrypted image of (e) and (f).

Fig. 16

Decrypted image of Salt and pepper noise at noise intensity of (a)1% (b) 3% (c) 10%; Decrypted image of speckle noise at noise intensity of (d)1%, (e) 3%, (f) 10%.

Fig. 17

(a) &(b) Cipher image rotated by ± 10° (c) & (d) Translated cipher image by ± 10 in both axes (e)&(f) Cipher image scaled by 0.8x,0.9x (g) & (h) Decrypted image of (a&b), (i) &(h) Decrypted image of (c &d), (k)&(l) Decrypted image of (e &f).

Fig. 18

(a) Original image (b) Decrypted image with change in key x, and (c) Decrypted image with a change in key y.