3 Hits in 1.9 sec

Ensemble Deep Learning Features for Real-World Image Steganalysis

2020 KSII Transactions on Internet and Information Systems  
The Alaska competition provides an opportunity to study the practical problems of real-world steganalysis.  ...  By separating the three colors channels for base model training and feature replacement strategy instead of simply merging features, the performance of the model ensemble is greatly improved.  ...  For example, we will try to use Siamese CNN [19] as a base model and combine different models and feature fusion strategies to improve our ensemble model.  ... 
doi:10.3837/tiis.2020.11.017 fatcat:jxxp7wophrfr5j47vivqws4fui

Multi-Contextual Design of Convolutional Neural Network for Steganalysis [article]

Brijesh Singh, Arijit Sur, Pinaki Mitra
2021 arXiv   pre-print
Most deep steganalysis classifiers usually extract noise residuals using high-pass filters as preprocessing steps and feed them to their deep model for classification.  ...  The model performance is further improved by incorporating the Self-Attention module to focus on the areas prone to steganalytic embedding.  ...  Steganalysis methods [12, 15, 19, 20, 22, 25] use different high-pass filters to suppress the image content and expose the noise content of an image.  ... 
arXiv:2106.10430v2 fatcat:uhl233bwrfbolds6ixxy77k5vq

Subsequent embedding in targeted image steganalysis: Theoretical framework and practical applications [article]

David Megías, Daniel Lerch-Hostalot
2022 arXiv   pre-print
Steganalysis is a collection of techniques used to detect whether secret information is embedded in a carrier using steganography.  ...  Although this solution has been applied in previous works, a theoretical basis for this strategy was missing.  ...  Acknowledgments We gratefully acknowledge the support of NVIDIA Corporation with the donation of an NVIDIA TITAN Xp GPU card that has been used in this work.  ... 
arXiv:2107.13862v3 fatcat:bwrbjywwbzg2vcshuxpgtm5tsm