ODIS-SD

[ODIS-SD-Optimized-Detection-of-Image-Steganography-in-Spatial-Domain

ODIS-SD

ODIS-SD focuses on enhancing the accuracy of steganalysis in identifying digital images holding confidential data. Current convolutional neural networks (CNNs) used for steganalysis have limitations in terms of accuracy. To address this, a novel approach is proposed, which optimizes the local features in the feature extraction stage. The performance of this approach is evaluated using the BOSSBase version 1.01 dataset and two adaptive steganography algorithms (WOW and S-UNIWARD) with payload sizes 0.2 and 0.4 bits per pixel (bpp). The results show an improvement of detection accuracy ranging from 2.1% to 3.6% compared to previous works.