Pixel-block-based Steganalysis Method for Hidden Data Location in Digital Images
Ntivuguruzwa Jean De La Croix, La Croix, Tohari Ahmad, Royyana Muslim Ijtihadie, N Croix, C Islamy, T Ahmad, T Ahmad, A Fatman, I Prayogi, T Ahmad, N Croix, P Maniriho, B Khosravi, B Khosravi, B Khosravi, K Nazarkardeh, N Croix, C Islamy, T Ahmad, I Thophile, N Croix, T Ahmad, A Ilham, T Ahmad, N Croix, P Maniriho, M Ntahobari, M Hossein, T Ahmad, N Croix, G Xie, J Ren, S Marshall, H Zhao, R Li, R Chen, N Croix, T Ahmad, C Rupa, S Shaikh, M Chinta, N Croix, T Ahmad, X Han, T Zhang, D Hu, Q Shen, S Zhou, X Liu, Y Fan, L Wang, B Pan, T Qiao, J Li, Y Chen, C Yang, T Qiao, X Luo, B Pan, Y Chen, X Wu, S Gupta, N Mohan, P Kaushal, J Fridrich, J Kodovsky, T Quach, Y Sun, H Zhang, T Zhang, R Wang, Q Liu, T Qiao, M Xu, N Zheng, T Filler, J Judas, J Fridrich, P Bas, T Filler, T Pevn
Abstract
Concealing confidential data within digital multimedia, also called steganography, has been crucial for safeguarding data.However, steganalysis techniques are best known for compromising these data concealment methods.The convolutional neural network (CNN) yielded better performance over the previously suggested techniques based on machine learning for steganalysis.Although the current approaches show promising outcomes, they encounter challenges in locating the secret data.This paper introduces a CNN architecture with a key idea to enhance the accuracy of identifying the precise hidden data location.Our method novelty is based on reducing image features and network parameters via customized feature allocation and convolutional kernels optimization and enhancing classification accuracy by optimizing the kernels combined with back-forward gradient descent.Our experimental results show improvements over the existing works with the accuracy under payload capacity of 0.4 bpp from 83.01% to 98.78% for S-UNIWARD and 87.19% to 97.93% for WOW.