In the midst of the progress in communication technology, most of the data used are available in internet and are in electronic form. Maintaining secrecy of data transmitted becomes a major issue. To provide information/data security, many techniques are available with each having their own merits and demerits. Steganography is one among them having the advantage of credulous hiding of data. Many steganography algorithms were developed to improve the robustness, level of security and capacity

more »

... the secret data hidden. In this proposed work Haar DWT is applied on the cover image and on the high frequency coefficients of the cover image, the compressed secret image is embedded at the neural network generated positions. From the stego image, the hidden image is extracted using neural network and decompressed to get the original information. In this pap er the robustness of the steganography is augmented with the help of frequency domain, the level of security is enhanced with the use of neural network and finally high capacity payload is achieved with the help of compression of secret data before embedding. Comparison of experimental results with the existing approaches shows that the proposed method shows better PSNR, invisibility of secret image and high capacity payload.