A Comparison of Ai Models To Detect Hidden Messages In Images

Steganography, the art of concealing secret information within seemingly innocuous cover media, poses significant challenges for digital security and forensic analysis. With the increasing use of digital images as carriers for hidden data, the need for efficient and accurate steganalysis techniques becomes imperative. This research compared several machine learning models including K-Nearest Neighbor, Gaussian, Multi-Layer Perceptron, Stochastic, Random Forest, aa non-pre-trained Convolutional Neural Network, and a pre-trained Convolutional Neural Network model called ResNet-18 in their effectiveness at detecting images that have a steganographic message embedded in it. The study found that the convolutional neural network was the best model in detecting steganographic content with 99% accuracy.