Contribution of neural networks in image steganography, watermarking and encryption
Abstract In front of information fishing, it is necessary to protect the information to be circulated over public channel. The goal of this paper is to propose an approach which inserts a logo inside of cover image and having three different possible results according to user's needs. The algor...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-02-01
|
| Series: | IET Image Processing |
| Online Access: | https://doi.org/10.1049/ipr2.12646 |
| _version_ | 1797937381533286400 |
|---|---|
| author | Maminiaina Alphonse Rafidison Sabine Harisoa Jacques Rafanantenana Andry Harivony Rakotomihamina Rajaonarison Faniriharisoa Maxime Toky Hajasoa Malalatiana Ramafiarisona |
| author_facet | Maminiaina Alphonse Rafidison Sabine Harisoa Jacques Rafanantenana Andry Harivony Rakotomihamina Rajaonarison Faniriharisoa Maxime Toky Hajasoa Malalatiana Ramafiarisona |
| author_sort | Maminiaina Alphonse Rafidison |
| collection | DOAJ |
| description | Abstract In front of information fishing, it is necessary to protect the information to be circulated over public channel. The goal of this paper is to propose an approach which inserts a logo inside of cover image and having three different possible results according to user's needs. The algorithm can perform steganography, watermarking or host encryption. This technique is based on singular value decomposition (SVD) and discrete wavelet transform (DWT) composed by three steps: pre‐processing, embedding and extraction/decryption. Reference matrix and reference parameter are the two factors deciding the operation and the intervention of neural networks makes the choice of their parameters easier according to the desired result. For measuring the steganography performance, the authors adopt the structural similarity index measure (SSIM) which calculates the similarity between original and watermarked image and peak signal‐to‐noise ratio. For watermarking, the normalized correlation (NC) coefficient investigates the correlation between the original watermark and the extracted watermark. Attacking the watermarked image with common attacks used by previous publication searchers, the value of SSIM and NC coefficient are closed to one. Regarding peak signal‐to‐noise ratio, the overall score is around 61.73 dB. The performance score is not negligible also for the encryption method. |
| first_indexed | 2024-04-10T18:43:41Z |
| format | Article |
| id | doaj.art-39b7b40589c2450d9013fafa70595a93 |
| institution | Directory Open Access Journal |
| issn | 1751-9659 1751-9667 |
| language | English |
| last_indexed | 2024-04-10T18:43:41Z |
| publishDate | 2023-02-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Image Processing |
| spelling | doaj.art-39b7b40589c2450d9013fafa70595a932023-02-01T11:19:25ZengWileyIET Image Processing1751-96591751-96672023-02-0117246347910.1049/ipr2.12646Contribution of neural networks in image steganography, watermarking and encryptionMaminiaina Alphonse Rafidison0Sabine Harisoa Jacques Rafanantenana1Andry Harivony Rakotomihamina2Rajaonarison Faniriharisoa Maxime Toky3Hajasoa Malalatiana Ramafiarisona4Telecommunication‐Automatic‐Signal‐Image‐Research Laboratory/Doctoral School in Science and Technology of Engineering and Innovation, University of Antananarivo Antananarivo MadagascarTelecommunication‐Automatic‐Signal‐Image‐Research Laboratory/Doctoral School in Science and Technology of Engineering and Innovation, University of Antananarivo Antananarivo MadagascarTelecommunication‐Automatic‐Signal‐Image‐Research Laboratory/Doctoral School in Science and Technology of Engineering and Innovation, University of Antananarivo Antananarivo MadagascarTelecommunication‐Automatic‐Signal‐Image‐Research Laboratory/Doctoral School in Science and Technology of Engineering and Innovation, University of Antananarivo Antananarivo MadagascarTelecommunication‐Automatic‐Signal‐Image‐Research Laboratory/Doctoral School in Science and Technology of Engineering and Innovation, University of Antananarivo Antananarivo MadagascarAbstract In front of information fishing, it is necessary to protect the information to be circulated over public channel. The goal of this paper is to propose an approach which inserts a logo inside of cover image and having three different possible results according to user's needs. The algorithm can perform steganography, watermarking or host encryption. This technique is based on singular value decomposition (SVD) and discrete wavelet transform (DWT) composed by three steps: pre‐processing, embedding and extraction/decryption. Reference matrix and reference parameter are the two factors deciding the operation and the intervention of neural networks makes the choice of their parameters easier according to the desired result. For measuring the steganography performance, the authors adopt the structural similarity index measure (SSIM) which calculates the similarity between original and watermarked image and peak signal‐to‐noise ratio. For watermarking, the normalized correlation (NC) coefficient investigates the correlation between the original watermark and the extracted watermark. Attacking the watermarked image with common attacks used by previous publication searchers, the value of SSIM and NC coefficient are closed to one. Regarding peak signal‐to‐noise ratio, the overall score is around 61.73 dB. The performance score is not negligible also for the encryption method.https://doi.org/10.1049/ipr2.12646 |
| spellingShingle | Maminiaina Alphonse Rafidison Sabine Harisoa Jacques Rafanantenana Andry Harivony Rakotomihamina Rajaonarison Faniriharisoa Maxime Toky Hajasoa Malalatiana Ramafiarisona Contribution of neural networks in image steganography, watermarking and encryption IET Image Processing |
| title | Contribution of neural networks in image steganography, watermarking and encryption |
| title_full | Contribution of neural networks in image steganography, watermarking and encryption |
| title_fullStr | Contribution of neural networks in image steganography, watermarking and encryption |
| title_full_unstemmed | Contribution of neural networks in image steganography, watermarking and encryption |
| title_short | Contribution of neural networks in image steganography, watermarking and encryption |
| title_sort | contribution of neural networks in image steganography watermarking and encryption |
| url | https://doi.org/10.1049/ipr2.12646 |
| work_keys_str_mv | AT maminiainaalphonserafidison contributionofneuralnetworksinimagesteganographywatermarkingandencryption AT sabineharisoajacquesrafanantenana contributionofneuralnetworksinimagesteganographywatermarkingandencryption AT andryharivonyrakotomihamina contributionofneuralnetworksinimagesteganographywatermarkingandencryption AT rajaonarisonfaniriharisoamaximetoky contributionofneuralnetworksinimagesteganographywatermarkingandencryption AT hajasoamalalatianaramafiarisona contributionofneuralnetworksinimagesteganographywatermarkingandencryption |