Hybrid Spread Spectrum Steganography using CNN

This paper uses a novel CNN based Spread spectrum steganography method. It is used to hide a secret image within a cover image using a pseudo-noise sequence in a spread spectrum manner, preserving the cover image’s original dimensions. Then we use CNN model for decryption to recover the secret image.A dataset with approximately 9700 images was used, the CNN model depicted excellent performance and overcame distortions introduced during embedding. The model was tested against evaluation parameters like PSNR, MSE and SSI.It achieved high Peak Signal-to-Noise Ratio (PSNR), low Mean Squared Error (MSE), and good Structural Similarity Index Measure (SSIM), confirming its robustness and effectiveness. This approach depicts the usage of a hybrid model to improve both traditional steganography and CNN based steganalysis, and is a strong tool for data protection.