Sensitivity of deep learning applied to spatial image steganalysis dataset

Dataset used on the paper "Sensitivity of deep learning applied to spatial image steganalysis", it comes from ”Break Our Steganographic System”: The Ins and Outs of Organizing BOSS from Patrick Bas, Tomas Filler, Tomas Pevny authorship , on the compress files cover.rar corresponds to cover images in pgm files, with the respective stego image from WOW and S-UNIWARD steganographic algoritms on Wow.rar and Suniward.rar respectively, and Wownpy.rar and Suninpy.rar contained npy files of cover-stego images divided on Train, Test and Validation sets.本数据集用于论文《面向空域图像隐写分析的深度学习敏感性分析》（*Sensitivity of deep learning applied to spatial image steganalysis*），该数据集源自Patrick Bas、Tomas Filler与Tomas Pevny合著的《破解我们的隐写系统：BOSS数据集的构建细节》（*Break Our Steganographic System: The Ins and Outs of Organizing BOSS*）。其中压缩包cover.rar内含PGM格式的载体图像；压缩包Wow.rar与Suniward.rar分别存储了采用WOW与S-UNIWARD两种隐写算法生成的隐写图像；Wownpy.rar与Suninpy.rar则存储了按训练集、测试集与验证集划分的载体-隐写图像对的npy（NumPy）格式文件。