Evaluating LSB and MSB Steganography in Retinal Fundus Images Through Image Quality Assessment and VGG19-Based Classification
DOI:
https://doi.org/10.30871/jaic.v10i3.13198Keywords:
Fundus Retina, LSB, MSB, Steganography, VGG19Abstract
The security of medical image data within electronic medical record systems has become a critical issue due to the increasing threat of health data breaches. Steganography is a promising technique for protecting patient information by concealing secret data within medical images without significantly altering their visual appearance. However, the application of steganography to retinal fundus images, which carry high diagnostic value, has never been comprehensively evaluated in terms of image quality or its impact on artificial intelligence-based diagnostic model performance. This study compares Least Significant Bit (LSB) and Most Significant Bit (MSB) steganography methods applied to 3,200 retinal fundus images from the Retinal Fundus Multi-disease Image Dataset (RFMiD) dataset across four payload levels (0.1-0.4 bpp), evaluated using PSNR, SNR, SSIM, and FSIM for image quality, and VGG19 classification accuracy and AUC for diagnostic impact. Results show LSB achieves substantially superior image quality (PSNR: 59.97-65.93 dB; SNR: 49.34-55.30 dB; SSIM: 0.9981-0.9997; FSIM: 0.9999-1.0000) compared to MSB (PSNR: 12.98-18.99 dB; SNR: 2.35-8.37 dB; SSIM: 0.5979-0.9003; FSIM: 0.5342-0.7500), while VGG19 classification accuracy remains stable for both methods (LSB: 0.8938-0.9000; MSB: 0.8953-0.9031) with a maximum difference of 0.62% from baseline. This study demonstrates that LSB is the more appropriate steganography method for retinal fundus images, delivering superior visual quality while preserving VGG19 diagnostic capability.
Downloads
References
[1] E. T. Ardianto, Sabran, dan L. Nurjanah, “Analisis Aspek Keamanan Data Pasien dalam Implementasi Rekam Medis Elektronik di Rumah Sakit X,” Jul 2024.
[2] M. Fitriyasari, “Tren dan Inovasi dalam Image Hiding untuk Keamanan Informasi Medis: Tinjauan Literatur,” RIGGS, vol. 4, no. 2, hlm. 2375–2381, Jun 2025, doi: 10.31004/riggs.v4i2.854.
[3] Z. Li, M. Xu, X. Yang, Y. Han, dan J. Wang, “A Multi-Label Detection Deep Learning Model with Attention-Guided Image Enhancement for Retinal Images,” Micromachines, vol. 14, no. 3, hlm. 705, Mar 2023, doi: 10.3390/mi14030705.
[4] T. Aswathi, T. R. Swapna, dan S. Padmavathi, “Transfer Learning approach for grading of Diabetic Retinopathy,” J. Phys.: Conf. Ser., vol. 1767, no. 1, hlm. 012033, Feb 2021, doi: 10.1088/1742-6596/1767/1/012033.
[5] R. Taj, F. Tao, S. Kanwal, A. Almogren, dan A. U. Rehman, “A SURF and SVD-based robust zero-watermarking for medical image integrity,” PLoS ONE, vol. 19, no. 9, hlm. e0307619, Sep 2024, doi: 10.1371/journal.pone.0307619.
[6] Ramyashree, P. S. Venugopala, S. Raghavendra, dan V. S. Kubihal, “Enhancing Secure Medical Data Communication Through Integration of LSB and DCT for Robust Analysis in Image Steganography,” IEEE Access, vol. 13, hlm. 1566–1580, 2025, doi: 10.1109/ACCESS.2024.3522957.
[7] S. Rahman, J. Uddin, H. U. Khan, H. Hussain, A. A. Khan, dan M. Zakarya, “A Novel Steganography Technique for Digital Images Using the Least Significant Bit Substitution Method,” IEEE Access, vol. 10, hlm. 124053–124075, 2022, doi: 10.1109/ACCESS.2022.3224745.
[8] S. Rustad, D. R. I. M. Setiadi, A. Syukur, dan P. N. Andono, “Inverted LSB image steganography using adaptive pattern to improve imperceptibility,” Journal of King Saud University - Computer and Information Sciences, vol. 34, no. 6, hlm. 3559–3568, Jun 2022, doi: 10.1016/j.jksuci.2020.12.017.
[9] S. N. M. Al‐Faydi, S. K. Ahmed, dan H. N. Y. Al‐Talb, “Improved LSB image steganography with high imperceptibility based on cover‐stego matching,” IET Image Processing, vol. 17, no. 7, hlm. 2072–2082, Mei 2023, doi: 10.1049/ipr2.12773.
[10] M. Z. Ali, O. Riaz, H. M. Hasnain, W. Sharif, T. Ali, dan G. S. Choi, “Elevating Image Steganography: A Fusion of MSB Matching and LSB Substitution for Enhanced Concealment Capabilities,” CMC, vol. 79, no. 2, hlm. 2923–2943, 2024, doi: 10.32604/cmc.2024.049139.
[11] M. Ilham dan C. Kirana, “Perbandingan Kualitas Citra Grayscale Steganografi Metode LSB dan DCT Berdasarkan PSNR dan SSIM,” vol. 7, no. 2, 2026, doi: https://doi.org/10.24076/joism.2026v7i2.2492.
[12] M. A. K. Raiaan dkk., “A Lightweight Robust Deep Learning Model Gained High Accuracy in Classifying a Wide Range of Diabetic Retinopathy Images,” IEEE Access, vol. 11, hlm. 42361–42388, 2023, doi: 10.1109/ACCESS.2023.3272228.
[13] K. A. Saputra dkk., “Multi-Disease Retinal Classification Using EfficientNet-B3 and Targeted Albumentations: A Benchmark on Kaggle Retinal Fundus Images Dataset,” SinkrOn, vol. 10, no. 1, hlm. 232–241, Jan 2026, doi: 10.33395/sinkron.v10i1.15530.
[14] N. Sengar, R. C. Joshi, M. K. Dutta, dan R. Burget, “EyeDeep-Net: a multi-class diagnosis of retinal diseases using deep neural network,” Neural Comput & Applic, vol. 35, no. 14, hlm. 10551–10571, Mei 2023, doi: 10.1007/s00521-023-08249-x.
[15] E. S. Kumar dan C. S. Bindu, “MDCF: Multi-Disease Classification Framework on Fundus Image Using Ensemble CNN Models,” vol. 40, no. 09, hlm. 35–45, 2021, doi: 10.17605/OSF.IO/ZHA9C.
[16] P. U. Pandey dkk., “Ensemble of deep convolutional neural networks is more accurate and reliable than board-certified ophthalmologists at detecting multiple diseases in retinal fundus photographs,” Br J Ophthalmol, vol. 108, no. 3, hlm. 417–423, Mar 2024, doi: 10.1136/bjo-2022-322183.
[17] E. Ho dkk., “Deep Ensemble Learning for Retinal Image Classification,” Trans. Vis. Sci. Tech., vol. 11, no. 10, hlm. 39, Okt 2022, doi: 10.1167/tvst.11.10.39.
[18] M. Kim, Y. Cho, H. Park, dan G. Qu, “ASIGM: An Innovative Adversarial Stego Image Generation Method for Fooling Convolutional Neural Network-Based Image Steganalysis Models,” Electronics, vol. 14, no. 4, hlm. 764, Feb 2025, doi: 10.3390/electronics14040764.
[19] S. Pachade dkk., “Retinal Fundus Multi-Disease Image Dataset (RFMiD): A Dataset for Multi-Disease Detection Research,” Data, vol. 6, no. 2, hlm. 14, Feb 2021, doi: 10.3390/data6020014.
[20] Pooja Belagali, Dr. V. R. Udupi, “Robust Image Steganography Based on Hybrid Edge Detection,” tjjpt, vol. 44, no. 3, hlm. 1509–1521, Okt 2023, doi: 10.52783/tjjpt.v44.i3.531.
[21] N. A. Mohsin dan H. A. Alameen, “A Hybrid Method for Payload Enhancement in Image Steganography Based on Edge Area Detection,” Cybernetics and Information Technologies, vol. 21, no. 3, hlm. 97–107, Sep 2021, doi: 10.2478/cait-2021-0032.
[22] Ramyashree, P. S. Venugopala, S. Raghavendra, dan B. Ashwini, “CrypticCare: A Strategic Approach to Telemedicine Security Using LSB and DCT Steganography for Enhancing the Patient Data Protection,” IEEE Access, vol. 12, hlm. 101166–101183, 2024, doi: 10.1109/ACCESS.2024.3430546.
[23] S. Rahman dkk., “A novel and efficient digital image steganography technique using least significant bit substitution,” Sci Rep, vol. 15, no. 1, hlm. 107, Jan 2025, doi: 10.1038/s41598-024-83147-3.
[24] A. A. Karawia, “Medical image steganographic algorithm via modified LSB method and chaotic map,” IET Image Processing, vol. 15, no. 11, hlm. 2580–2590, Sep 2021, doi: 10.1049/ipr2.12246.
[25] Y. Sanjalawe, S. Al-E’mari, S. Fraihat, M. Abualhaj, dan E. Alzubi, “A deep learning-driven multi-layered steganographic approach for enhanced data security,” Sci Rep, vol. 15, no. 1, hlm. 4761, Feb 2025, doi: 10.1038/s41598-025-89189-5.
[26] M. Naufal, H. A. Azies, F. A. Zami, dan R. M. Brilianto, “Optimizing Driver Drowsiness Detection: Evaluating CLAHE and AHE Enhancement Techniques,” vol. 15.
[27] Y. Al Najjar, “Comparative Analysis of Image Quality Assessment Metrics: MSE, PSNR, SSIM and FSIM,” IJSR, vol. 13, no. 3, hlm. 110–114, Mar 2024, doi: 10.21275/SR24302013533.
[28] A. M. Almhlbdi, N. D. Altowairqi, A. O. Alshutayri, dan R. K. Qarout, “Deep Learning-Based Multi-Class Detection of LSB Steganography in Digital Images,” IEEE Access, vol. 13, hlm. 191543–191553, 2025, doi: 10.1109/ACCESS.2025.3628784.
[29] M. F. Alexander, V. R. T. Adrian, L. R. Esprayenduo, dan M. Naufal, “Klasifikasi Penyakit Mata Menggunakan Random Forest Dengan Optimasi Hyperparameter RandomSearchCV,” vol. 5, no. 2, 2026.
[30] L. Lerch dkk., “DreamOn: a data augmentation strategy to narrow the robustness gap between expert radiologists and deep learning classifiers,” Front Radiol, vol. 4, hlm. 1420545, 2024, doi: 10.3389/fradi.2024.1420545.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Gilang Faturrahman, Muhammad Naufal, Wahyu Aji Eko Prabowo, Sindhu Rakasiwi
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
