StegoLab

In StegoLab you can find implementations of different techniques used in steganography, steganalysis and watermarking.

Steganography

Codes:

• Binary Hamming codes: Implementation in Python of information hiding using matrix embedding techniques with binary Hamming codes. You can find more information in the article “Binary Hamming Codes in Steganography”.

• Ternary Hamming codes: Implementation in Python of information hiding using matrix embedding techniques with ternary Hamming codes. Ternary codes allow a higher capacity than binary codes for the same level of media distortion.You can find more information in the article “Ternary Hamming Codes in Steganography”.

• Wet Paper Codes: A Python implementación of the steganography embedding technique presented in the paper “Writing on Wet Paper” by Jessica Fridrich, Miroslav Goljan, Petr Lisoněk and David Soukal

• Syndrome Trellis Codes: A Python implementation of the steganography embedding technique presented in the paper “Minimizing embedding impact in steganography using trellis-coded quantization” by Tomáš Filler, Jan Judas and Jessica Fridrich.

Cost functions:

• S-UNIWARD: A Python implementation of the cost function used in steganography for hidding information into bitmap images, proposed in the paper “Universal Distortion Function for Steganography in an Arbitrary Domain” by Vojtěch Holub, Jessica Fridrich and Tomáš Denemark.

• J-UNIWARD: A Python implementation of the cost function used in steganography for hidding information into JPEG images, proposed in the paper “Universal Distortion Function for Steganography in an Arbitrary Domain” by Vojtěch Holub, Jessica Fridrich and Tomáš Denemark. Includes fast implementations using Numba and implementations that decompress into RGB without rounding the pixel values.

• J-UNIWARD + Cost Polarization: A Python implementation of the cost polarization technique using the Wiener filter for J-UNIWARD proposed in the article JPEG Steganography With Estimated Side-Information by Weixiang Li, Kejiang Chen, Weiming Zhang, Hang Zhou, Yaofei Wang and Nenghai Yu.

• RBV: A Python implementation of the cost function used in steganography for hidding information into JPEG images, proposed in the paper “Distortion function based on residual blocks for JPEG steganography” by Qingde Wei, Zhaoxia Yin, Zichi Wang and Xinpeng Zhang.

• HILL: A Python implementation of the cost function used in steganography for hiding information into bitmap images, proposed in the paper “A New Cost Function for Spatial Image Steganography” by Bin Li, Ming Wang, Jiwu Huang and Xiaolong Li.

Steganalysis

• ATS attack: Python implementation of the ATS attack, an unsupervised steganalysis technique presented in the article Unsupervised steganalysis based on artificial training sets [arXiv] by Daniel Lerch-Hostalot and David Megías.

• Calibration Attack: Implementation of the attack to F5 algorithm (JPEG steganography) proposed in the paper Steganalysis of JPEG Images: Breaking the F5 Algorithm by Jessica Fridrich, Miroslav Goljan and Dorin Hogea.

• pyEC: Python Interface to the Matlab version of Ensemble Classifiers for Steganalysis, presented in the paper “Ensemble Classifiers for Steganalysis of Digital Media” by Jan Kodovský, Jessica Fridrich and Vojtěch Holub.

Watermarking

• E-Blind: Blind embedding and Linear Correlation Detection. This method simply adds a pattern to the image. The method is based on “System 1” proposed in the book Digital Watermarking and Steganography de I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich y T. Kalker.

• E-Fixed-LC: Fixed Linear Correlation Embedder (informed embedding) and Linear Correlation Detection. This method adjusts the strength of the watermark to ensure that the watermarked image has a specified linear correlation. The method is based on “System 2” proposed in the book Digital Watermarking and Steganography de I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich y T. Kalker.

• E-blk-Blind: Block-based blind embedding and correlation coefficient detection. This method works in a similar way to the E-Blind method. The method is based on “System 3” proposed in the book Digital Watermarking and Steganography de I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich y T. Kalker.

• E-Simple-8: 8-bit blind embedder and 8-bit detector. This method is a version of E-Blind modified to embed 8-bits. The method is based on “System 4” proposed in the book Digital Watermarking and Steganography de I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich y T. Kalker.

• E-Trellis-8: Trellis-coding embedder and Viterbi detector. This method embeds 8-bit messages using trellis-coded modulation. The method is based on “System 4” proposed in the book Digital Watermarking and Steganography de I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich y T. Kalker.