Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise

Qinglong Huang; Haoran Li; Yong Liao; Yanbin Hao; Pengyuan Zhou

doi:10.1007/978-3-031-72335-3_22

Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise

Qinglong Huang, Haoran Li, Yong Liao^*, Yanbin Hao, Pengyuan Zhou^*

^*Corresponding author for this work

Department of Electrical and Computer Engineering - Communication, Control and Automation

Research output: Contribution to book/anthology/report/proceeding › Article in proceedings › Research › peer-review

1 Citation (Scopus)

Abstract

Neural Radiance Field (NeRF) has been proposed as an innovative advancement in 3D reconstruction techniques. However, little research has been conducted on the issues of information confidentiality and security to NeRF, such as steganography. Existing NeRF steganography solutions have shortcomings in low steganography quality, model weight damage, and limited amount of steganographic information. This paper proposes Noise-NeRF, a novel NeRF steganography method employing Adaptive Pixel Selection strategy and Pixel Perturbation strategy to improve the quality and efficiency of steganography via trainable noise. Extensive experiments validate the state-of-the-art performances of Noise-NeRF on both steganography quality and rendering quality, as well as effectiveness in super-resolution image steganography.

Original language	English
Title of host publication	Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings
Editors	Michael Wand, Kristína Malinovská, Jürgen Schmidhuber, Igor V. Tetko
Number of pages	15
Publisher	Springer
Publication date	2024
Pages	320-334
ISBN (Print)	9783031723346
DOIs	https://doi.org/10.1007/978-3-031-72335-3_22
Publication status	Published - 2024
Event	33rd International Conference on Artificial Neural Networks, ICANN 2024 - Lugano, Switzerland Duration: 17 Sept 2024 → 20 Sept 2024

Conference

Conference	33rd International Conference on Artificial Neural Networks, ICANN 2024
Country/Territory	Switzerland
City	Lugano
Period	17/09/2024 → 20/09/2024

Series	Lecture Notes in Computer Science
Volume	15017 LNCS
ISSN	0302-9743

Keywords

implicit neural representation
neural radiation fields
steganography

Access to Document

10.1007/978-3-031-72335-3_22

Cite this

Huang, Q., Li, H., Liao, Y., Hao, Y., & Zhou, P. (2024). Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise. In M. Wand, K. Malinovská, J. Schmidhuber, & I. V. Tetko (Eds.), Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings (pp. 320-334). Springer. https://doi.org/10.1007/978-3-031-72335-3_22

Huang, Qinglong ; Li, Haoran ; Liao, Yong et al. / Noise-NeRF : Hide Information in Neural Radiance Field Using Trainable Noise. Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings. editor / Michael Wand ; Kristína Malinovská ; Jürgen Schmidhuber ; Igor V. Tetko. Springer, 2024. pp. 320-334 (Lecture Notes in Computer Science, Vol. 15017 LNCS).

@inproceedings{9bb6b85cda5749168ea858ee4c243271,

title = "Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise",

abstract = "Neural Radiance Field (NeRF) has been proposed as an innovative advancement in 3D reconstruction techniques. However, little research has been conducted on the issues of information confidentiality and security to NeRF, such as steganography. Existing NeRF steganography solutions have shortcomings in low steganography quality, model weight damage, and limited amount of steganographic information. This paper proposes Noise-NeRF, a novel NeRF steganography method employing Adaptive Pixel Selection strategy and Pixel Perturbation strategy to improve the quality and efficiency of steganography via trainable noise. Extensive experiments validate the state-of-the-art performances of Noise-NeRF on both steganography quality and rendering quality, as well as effectiveness in super-resolution image steganography.",

keywords = "implicit neural representation, neural radiation fields, steganography",

author = "Qinglong Huang and Haoran Li and Yong Liao and Yanbin Hao and Pengyuan Zhou",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.; 33rd International Conference on Artificial Neural Networks, ICANN 2024 ; Conference date: 17-09-2024 Through 20-09-2024",

year = "2024",

doi = "10.1007/978-3-031-72335-3_22",

language = "English",

isbn = "9783031723346",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "320--334",

editor = "Michael Wand and Krist{\'i}na Malinovsk{\'a} and J{\"u}rgen Schmidhuber and Tetko, {Igor V.}",

booktitle = "Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings",

address = "Netherlands",

}

Huang, Q, Li, H, Liao, Y, Hao, Y & Zhou, P 2024, Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise. in M Wand, K Malinovská, J Schmidhuber & IV Tetko (eds), Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings. Springer, Lecture Notes in Computer Science, vol. 15017 LNCS, pp. 320-334, 33rd International Conference on Artificial Neural Networks, ICANN 2024, Lugano, Switzerland, 17/09/2024. https://doi.org/10.1007/978-3-031-72335-3_22

Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise. / Huang, Qinglong; Li, Haoran; Liao, Yong et al.
Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings. ed. / Michael Wand; Kristína Malinovská; Jürgen Schmidhuber; Igor V. Tetko. Springer, 2024. p. 320-334 (Lecture Notes in Computer Science, Vol. 15017 LNCS).

Research output: Contribution to book/anthology/report/proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - Noise-NeRF

T2 - 33rd International Conference on Artificial Neural Networks, ICANN 2024

AU - Huang, Qinglong

AU - Li, Haoran

AU - Liao, Yong

AU - Hao, Yanbin

AU - Zhou, Pengyuan

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

PY - 2024

Y1 - 2024

N2 - Neural Radiance Field (NeRF) has been proposed as an innovative advancement in 3D reconstruction techniques. However, little research has been conducted on the issues of information confidentiality and security to NeRF, such as steganography. Existing NeRF steganography solutions have shortcomings in low steganography quality, model weight damage, and limited amount of steganographic information. This paper proposes Noise-NeRF, a novel NeRF steganography method employing Adaptive Pixel Selection strategy and Pixel Perturbation strategy to improve the quality and efficiency of steganography via trainable noise. Extensive experiments validate the state-of-the-art performances of Noise-NeRF on both steganography quality and rendering quality, as well as effectiveness in super-resolution image steganography.

AB - Neural Radiance Field (NeRF) has been proposed as an innovative advancement in 3D reconstruction techniques. However, little research has been conducted on the issues of information confidentiality and security to NeRF, such as steganography. Existing NeRF steganography solutions have shortcomings in low steganography quality, model weight damage, and limited amount of steganographic information. This paper proposes Noise-NeRF, a novel NeRF steganography method employing Adaptive Pixel Selection strategy and Pixel Perturbation strategy to improve the quality and efficiency of steganography via trainable noise. Extensive experiments validate the state-of-the-art performances of Noise-NeRF on both steganography quality and rendering quality, as well as effectiveness in super-resolution image steganography.

KW - implicit neural representation

KW - neural radiation fields

KW - steganography

UR - http://www.scopus.com/inward/record.url?scp=85205881756&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-72335-3_22

DO - 10.1007/978-3-031-72335-3_22

M3 - Article in proceedings

AN - SCOPUS:85205881756

SN - 9783031723346

T3 - Lecture Notes in Computer Science

SP - 320

EP - 334

BT - Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings

A2 - Wand, Michael

A2 - Malinovská, Kristína

A2 - Schmidhuber, Jürgen

A2 - Tetko, Igor V.

PB - Springer

Y2 - 17 September 2024 through 20 September 2024

ER -

Huang Q, Li H, Liao Y, Hao Y, Zhou P. Noise-NeRF: Hide Information in Neural Radiance Field Using Trainable Noise. In Wand M, Malinovská K, Schmidhuber J, Tetko IV, editors, Artificial Neural Networks and Machine Learning – ICANN 2024 - 33rd International Conference on Artificial Neural Networks, Proceedings. Springer. 2024. p. 320-334. (Lecture Notes in Computer Science, Vol. 15017 LNCS). doi: 10.1007/978-3-031-72335-3_22