Robust secret image sharing scheme with improved anti-noise capability

Abstract

Secret image sharing (SIS) is a promising image protection technology boasting high security and loss tolerance. However, shadow images generated by SIS are susceptible to noise contamination during real-world transmission, storage, and processing, resulting in data loss when recovering the secret image. Therefore, improving the robustness of SIS systems against noise is crucial. Unfortunately, there is currently limited research on this topic, and existing schemes require improvement in the anti-noise performance. Under the premise of no pixel expansion, this paper proposes a robust SIS scheme that can effectively resist many types of shadow image noise, single or compound. We have achieved the organic fusion of SIS and Reed–Solomon coding technology by elaborately designing many critical factors, such as the shortened Reed–Solomon codes, the construction and utilization method of pixel-blocks, and other prerequisites. The proposed scheme is particularly effective in resisting burst noise, such as salt-and-pepper noise, with lossless recovery possible when the noise density d≤0.01 and high-quality recovery when d<0.06. Compared with existing schemes, it has many advantages, such as stronger anti-noise capability, various applicable noise types, stable anti-noise performance, and no pixel expansion. Analyses and illustrations are conducted to validate the proposed scheme.