Dual-image based high image quality reversible hiding scheme with multiple folding zones

A dual-image technique is often used in reversible data hiding. In this technique, secret information is concealed within a host image to generate two stego-images. In 2015, Lu et al. proposed a center-folding strategy to enhance the performance of the reversible hiding scheme. In their scheme, the value range of the secret data is divided into two zones. The medium value is subtracted from each value of the value range to generate the folded value. The folded value is smaller than that of the original value such that the image quality of the stego-image can be increased. However, there is still room for improvement. This study proposes a novel method to further narrow down the folded value. The proposed scheme divides the value range into several zones instead of only two zones. Furthermore, the scheme re-encodes the secret information according to the number of zones and the maximum distortion between the stego-pixel and the host pixel. The re-encoded value is then separated into two parts and embedded in an image. Through this process, image quality can be significantly improved. The proposed scheme uses different zone sizes and the maximum distortion to control the hiding rate and image quality. The hiding rates of the proposed scheme with a zone size set at 12 and maximum distortion of each pixel set at 5 is 1.75 bit per pixel (bpp), and the image quality is 43.22 dB, which is higher than that of Lu et al.’s scheme, Lee et al.’s scheme, Chang’s scheme and Wang’s scheme. Using a similar hiding rate, the image quality of the proposed scheme has a higher embedding performance than that of the schemes of Xue et al., Arham et al., and Zhang et al.