| LSSD : Large Scale
Steganalysis Database is a database of 2 million JPEG (color or greyscale) 256x256 images. Webpage published in September 2020 and "alive" in January 2021. Here is the material for: - downloading the database thanks script: LSSD_download_script.sh; Otherwise, direct access via the dedicated website is possible: http://lssd.lirmm.fr/ Note that a GitHub is also available (https://github.com/Yiouki/download_lssd) - description of the LSSD folder hierarchy: DB_structure.pdf - re-developing the RAW images (inspired from Alaska#1 script): jpeg_base_generator.zip, - using the LC network (Low Complexity - paper - github) CNN: low_complexity_tf.zip. Any use of this database must be mentioned by citing the following papers: ' LSSD: a Controlled Large JPEG Image
Database for Deep-Learning-based Steganalysis ”into the Wild”
', Hugo RUIZ, Mehdi
YEDROUDJ,
Marc
CHAUMONT, Frédéric COMBY, Gérard SUBSOL, Proceeding of the 25th International Conference on Pattern Recognition, ICPR'2021, Worshop on MultiMedia
FORensics in the WILD, MMForWILD'2021, Virtual Conference due to Covid (formerly
Milan, Italy),
Lecture Notes in Computer Science, LNCS 12666, Springer,
Book: Pattern Recognition and Information Forensics, Chapter No: 38, Pages 470-483, Chapter DOI: 10.1007/978-3-030-68780-9_38,
January 10-15, 2021, 14 pages, https://iplab.dmi.unict.it/mmforwild/. paper, slides, video-lirmm.
Alaska#2: "The ALASKA Steganalysis Challenge: A First Step Towards Steganalysis". Cogranne, R., Giboulot, Q., Bas, P. In:
Proceedings of the ACM Workshop on Information Hiding and Multimedia
Security. pp. 125–137. IH&MMSec’2019, Paris, France (Jul 2019).
StegoApp: "StegoAppDB: a Steganography Apps Forensics Image Database".
Newman, J., Lin, L., Chen, W., Reinders, S., Wang, Y., Wu, M., Guan, Y.
In: Proceedings ofMedia Watermarking, Security, and Forensics,
MWSF’2019, Part of IS&T International Symposium on Electronic
Imaging, EI’2019. Ingenta, Burlingame, California,USA (Jan 2019).
BOSS: "Break Our Steganographic System: The Ins and Outs of Organizing BOSS".
Bas, P., Filler, T., Pevny, T. In: Proceedings of 13th International
Conference on In-formation Hiding, IH’2011. Lecture Notes in Computer
Science, Springer, vol. 6958,pp. 59–70. Prague, Czech Republic (May
2011).
Raise : "RAISE – A Raw ImagesDataset for Digital Image Forensics".
Dang-Nguyen, D.T., Pasquini, C., Conotter, V., Boato, G. In:
Proceedings of ACM Multimedia Systems.Portland, Oregon (March 2015).
Dresden : "The ‘Dresden Image Database’ for Benchmarking DigitalImage Forensics".
Gloe, T., Bohme, R.: In: Proceedings of the
25th Symposium On Applied Computing (ACM SAC 2010). vol. 2, pp.
1585–1591 (2010).
 Fig: Diagram of the RAW image development process
The LSSD database contains: - The RAW_BASE: We gathered different sources of RAW images, and you can download them separately thanks to the script (LSSD_download_script.sh). Note that we decided to store all those RAW images on our Website (http://lssd.lirmm.fr/RAW/), in order to archive the different sources and help the scientific community. The sources are ALASKA 2, BOSS, Dresden, RAISE, Stego App, and Wesaturate, (some of them are accessible on the associated website) :
- LSSD_JPEG_256x256_2M: The LSSD (you can select the color or grey-level image database) database obtained by developing the RAW_BASE (see the paper for explanations). There are 2 millions of images either in JPEG or in MAT format (MAT = decompressed and in double precision values). Command to download JPEG: sh LSSD_download_script.sh -b LSSD_2M -t JPEG -c Color (or Gray) -n Cover Note: command to download MAT: sh LSSD_download_script.sh -b LSSD_2M -t MAT -c Color (or Gray) -n Cover. - LSSD_JPEG_256x256_1M: Images from LSSD_JPEG_256x256_1M are all included in LSSD_JPEG_256x256_2M. Command to download JPEG or MAT: sh LSSD_download_script.sh -b LSSD_1M -t JPEG (or MAT) -c Color (or Gray) -n Cover. - LSSD_JPEG_256x256_500K: Images from LSSD_JPEG_256x256_500K are all included in LSSD_JPEG_256x256_1M. Command to download JPEG or MAT: sh LSSD_download_script.sh -b LSSD_500k -t JPEG (or MAT) -c Color (or Gray) -n Cover. -LSSD_JPEG_256x256_100K: Images from LSSD_JPEG_256x256_100K are all included in LSSD_JPEG_256x256_500K. Command to download JPEG or MAT: sh LSSD_download_script.sh -b LSSD_100k -t JPEG (or MAT) -c Color (or Gray) -n Cover. - LSSD_JPEG_256x256_50K: Images from LSSD_JPEG_256x256_50K are all included in LSSD_JPEG_256x256_100K. Command to download JPEG or MAT: sh LSSD_download_script.sh -b LSSD_50k -t JPEG (or MAT) -c Color (or Gray) -n Cover. - LSSD_JPEG_256x256_10K: Images from LSSD_JPEG_256x256_10K are all included in LSSD_JPEG_256x256_50K. Command to download JPEG or MAT: sh LSSD_download_script.sh -b LSSD_10k -t JPEG (or MAT) -c Color (or Gray) -n Cover. - LSSD_JPEG_256x256_TEST: This database is disjoint from the LSSD (The RAW which have been used in order to generate this test database are different from those used in order to generate the LSSD_JPEG_256x256_2M). Command to download JPEG or MAT: sh LSSD_download_script.sh -b TST_100k -t JPEG (or MAT) -c Color (or Gray) -n Cover. -- For a rapid use, we also gave below the stego database with an embedding with JUNIWARD at 0.2 bpnzacs: - LSSD_JPEG_256x256_2M_STEGO. Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b LSSD_2M -t JPEG (or MAT) -c Gray -n Stego_P02. - LSSD_JPEG_256x256_1M_STEGO. Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b LSSD_1M -t JPEG (or MAT) -c Gray -n Stego_P02. - LSSD_JPEG_256x256_500K_STEGO. Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b LSSD_500k -t JPEG (or MAT) -c Gray -n Stego_P02. - LSSD_JPEG_256x256_100K_STEGO. Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b LSSD_100k -t JPEG (or MAT) -c Gray -n Stego_P02. - LSSD_JPEG_256x256_50K_STEGO. Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b LSSD_50k -t JPEG (or MAT) -c Gray -n Stego_P02. - LSSD_JPEG_256x256_10K_STEGO: Images from LSSD_JPEG_256x256_10K are all included in LSSD_JPEG_256x256_50K. Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b LSSD_50k -t JPEG (or MAT) -c Gray -n Stego_P02. -- - LSSD_JPEG_256x256_TEST_STEGO: Command to download JPEG or MAT (only in grey-level): sh LSSD_download_script.sh -b TST_100k -t JPEG (or MAT) -c Gray -n Stego_P02. -- Few words on the project: The objective of our project is to study the behavior of steganalysis with huge learning bases and in controlled conditions. The original images are in RAW format. This is
the information perceived by the
camera physical sensor when the image is captured. It gives large files
which are quite rare to find on the Internet. We give you access
to all the RAW images, sorted according to their origin. The given scripts are the ones we used throughout our project. One is especially for you: it will allow you to download any LSSD database part. One of the scripts allows you to develop JPEG images from RAW images. And the last one allows the Low Complexity neural network to be leart on the selected database. HOW TO CITE the LSSD database:If researchers use this free-access database, they are required to cite those papers: ' LSSD: a Controlled Large JPEG Image
Database for Deep-Learning-based Steganalysis ”into the Wild”
', Hugo RUIZ, Mehdi
YEDROUDJ,
Marc
CHAUMONT, Frédéric COMBY, Gérard SUBSOL, Proceeding of the 25th International Conference on Pattern Recognition, ICPR'2021, Worshop on MultiMedia
FORensics in the WILD, MMForWILD'2021, Virtual Conference due to Covid (formerly
Milan, Italy),
Lecture Notes in Computer Science, LNCS 12666, Springer,
Book: Pattern Recognition and Information Forensics, Chapter No: 38, Pages 470-483, Chapter DOI: 10.1007/978-3-030-68780-9_38,
January 10-15, 2021, 14 pages, https://iplab.dmi.unict.it/mmforwild/. paper, slides, video-lirmm.
Alaska#2: "The ALASKA Steganalysis Challenge: A First Step Towards Steganalysis". Cogranne, R., Giboulot, Q., Bas, P. In:
Proceedings of the ACM Workshop on Information Hiding and Multimedia
Security. pp. 125–137. IH&MMSec’2019, Paris, France (Jul 2019).
StegoApp: "StegoAppDB: a Steganography Apps Forensics Image Database".
Newman, J., Lin, L., Chen, W., Reinders, S., Wang, Y., Wu, M., Guan, Y.
In: Proceedings ofMedia Watermarking, Security, and Forensics,
MWSF’2019, Part of IS&T International Symposium on Electronic
Imaging, EI’2019. Ingenta, Burlingame, California,USA (Jan 2019).
BOSS: "Break Our Steganographic System: The Ins and Outs of Organizing BOSS".
Bas, P., Filler, T., Pevny, T. In: Proceedings of 13th International
Conference on In-formation Hiding, IH’2011. Lecture Notes in Computer
Science, Springer, vol. 6958,pp. 59–70. Prague, Czech Republic (May
2011).
Raise : "RAISE – A Raw ImagesDataset for Digital Image Forensics".
Dang-Nguyen, D.T., Pasquini, C., Conotter, V., Boato, G. In:
Proceedings of ACM Multimedia Systems.Portland, Oregon (March 2015).
Dresden : "The ‘Dresden Image Database’ for Benchmarking DigitalImage Forensics".
Gloe, T., Bohme, R.: In: Proceedings of the
25th Symposium On Applied Computing (ACM SAC 2010). vol. 2, pp.
1585–1591 (2010). Related work: ' Analysis of the Scalability of a
Deep-Learning Network for Steganography “Into the Wild” ',
Hugo RUIZ,
Marc
CHAUMONT, Mehdi YEDROUDJ, Ahmed OULAD AMARA, Frédéric
COMBY, Gérard SUBSOL, Proceeding of the 25th International Conference on Pattern Recognition, ICPR'2021, Worshop on MultiMedia
FORensics in the WILD, MMForWILD'2021, Virtual Conference due to Covid (formerly
Milan, Italy), Lecture Notes in Computer Science, LNCS 12666, Springer,
Book: Pattern Recognition and Information Forensics, Chapter No: 36, Pages 439-452, Chapter DOI: 10.1007/978-3-030-68780-9_36,
January 10-15, 2021, 14 pages, https://iplab.dmi.unict.it/mmforwild/. paper. slides. video-lirmm, video-screencast-o-matic, video-youtube.
|
||||||||||||||||||||||||||||||||||||||||
