A Novel Image Steganographic Method based on Integer Wavelet Transformation and Particle Swarm Optimization

Abstract Image steganography is a technique of hiding secret data into a cover image and so as to prevent the intruders from accessing the secret data. The efficiency of image steganography techniques are usually evaluated based on perceptual transparency, payload capacity, security, temper resistance and computational costs. Though there has been significant improvement in related research over the decades, available steganographic techniques usually satisfy only a subset of these criteria. This paper presents a novel IWT (Integer Wavelet Transform) based steganography method using PSO (Particle Swarm Optimization) to find the optimal substitution matrix for converting secret data into their substituted forms. In the proposed method, optimal pixel adjustment process is used to improve perceptual transparency so that the obtained stego image has low distortion. The proposed method improves the security, imperceptibility, and robustness of the secret data by hiding them into the wavelet coefficients of an image. Thus, the paper provides a detailed study of the use of PSO in three different image steganographic methods based on (i) LSB (Least Significant Bit) substitution, (ii) DWT (Discrete Wavelet Transform), and (iii) IWT. Experiments are conducted using well-known benchmark images and results are comparatively analyzed. It is found that our proposed approach of PSO based IWT outperforms both PSO based LSB and PSO based DWT from the context of standard quality metrics, statistical analysis, security level estimation, payload capacity, imperceptibility etc. We have also provided a comparative overview of the existing data hiding methods including the proposed approach.