Computational Methods of Digital Steganography
УДК: 519.688
Abstract
The paper reviews current digital steganography techniques for photo and video content. It introduces a neural-network algorithm for covert data transmission in video streams. The architecture of computational methods blends 2D and 3D convolutions: the 2D layers preserve fine frame details, while the 3D layer averages features across three neighboring frames to suppress inter-frame «flicker». Data are embedded frame-by-frame via a sliding window, so only a trio of frames needs to reside in memory, enabling efficient streaming. During training, a differentiable distortion module is used to make the system robust against real-world rerecording noise. The composite loss combines masked MSE (embedding imperceptibility), binary cross-entropy (decoding accuracy), and a temporal inconsistency penalty. The proposed method with a payload of 256 bits per frame attains PSNR=31 dB, SSIM=0.92.
Downloads
References
Федосеев В.А., Митекин В.А. Теоретические основы стеганографии и цифровых водяных знаков. Самара: Изд-во Самарского ун-та, 2017. 130 с.
Johnson N.F., Katzenbeisser S. A Survey of Stegano-graphic Techniques // Information Hiding Techniques for Ste-ganography and Digital Watermarking / red. S.A. Katzenbeisser, F. P. Petitcolas. Boston: Artech House, 2000. P. 43-78.
Chan C.K., Cheng L.M. Hiding Data in Images by Simple LSB Substitution // Pattern Recognition. 2004. Vol. 37. № 3. P. 469-474. https://doi.org/10.1016/j.patcog.2003.08.007
Baluja S. Hiding Images in Plain Sight: Deep Ste-ganography // Advances in Neural Information Processing Systems 30 (NeurIPS 2017). Red Hook, NY: Curran Associates, Inc., 2017. P. 2069-2079.
Zhu J., Kaplan R., Johnson J., Li F.-F. HiDDeN: Hiding Data with Deep Networks // Computer Vision — ECCV 2018. Lecture Notes in Computer Science, vol. 11129. Cham: Springer, 2018. P. 682-697. https://doi.org/10.1007/978-3-030-01267-0_40
Bui T., Agarwal S., Yu N., Collomosse J. RoSteALS: Robust Steganography Using Autoencoder Latent Space // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW 2023). Los Alamitos, CA: IEEE Computer Society, 2023. P 933-942. https://doi.org/10.1109/CVPRW59228.2023.00100
Sobel I., Feldman G. A 3x3 Isotropic Gradient Operator for Image Processing. Stanford: Stanford Artificial Intelligence Laboratory, 1968. 10 p. ResearchGate. URL: https://www.researchgate.net/publication/239398674_A_3x3_Isotropic_ Gradient_Operator_for_Image_Processing (дата обращения: 09.08.2025).
Шустов С.А., Мещеряков РВ. Модель стеганографического встраивания информации в изображения методами глубокого обучения // Информационные процессы. 2025. Т. 25. № 1. С. 1-13. https://doi.org/10.53921/18195822_2025_25_1_1
Kingma D.P, Ba J.L. Adam: A Method for Stochastic Optimization // International Conference on Learning Representations (ICLR 2015), San Diego, CA. Ithaca, NY: Cornell University, 2015. ArXiv e-print. URL: https://arxiv.org/abs/1412.6980 (дата обращения: 09.08.2025).
Tancik M., Mildenhall B., Ng R. StegaStamp: Invisible Hyperlinks in Physical Photographs // Proceedings of the IEEE/ CVF Conference on Computer Vision and Pattern Recognition (CVPR 2020). Los Alamitos, CA: IEEE Computer Society, 2020. P. 2114-2123. https://doi.org/10.1109/CVPR42600.2020.00219
Copyright (c) 2026 Сергей Александрович Шустов, Роман Валерьевич Мещеряков
This work is licensed under a Creative Commons Attribution 4.0 International License.
Izvestiya of Altai State University is a golden publisher, as we allow self-archiving, but most importantly we are fully transparent about your rights.
Authors may present and discuss their findings ahead of publication: at biological or scientific conferences, on preprint servers, in public databases, and in blogs, wikis, tweets, and other informal communication channels.
Izvestiya of Altai State University allows authors to deposit manuscripts (currently under review or those for intended submission to Izvestiya of Altai State University) in non-commercial, pre-print servers such as ArXiv.
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 (CC BY 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).
