A Neural Network Based Image Steganography Method using Cyclic Chaos and Integer Wavelet Transform
Manoj Kumar, Tahera Hussaini · 2021 Asian Conference on Innovation in Technology (ASIANCON) · 2021
Steganography is the technique of hiding confidential information inside a cover in order to deceive attackers. Neural Networks (NN) become very popular and effective to train a model and get accurate result. In this article we used unsupervised NN to select and train model for cover image selection. Chaotic structures and transformation techniques like the integer wavelet transform in Steganography have good result due to minimum data loss. In this paper, we propose a cyclic chaos algorithm that produces Pseudorandom Number Generator (PRNG), use seeds for embedding purpose. We also used the Integer Wavelet Transform (IWT) to minimize data loss and return a high-quality encoded image. Our proposed system can embed colored image. Furthermore, evident and easily understandable as well as executable. We found that our approach outperforms other recent steganography approaches in terms of improved embedding time, reduced complexity, and improved decoded image quality with increased hiding capability.