A novel real-time and progressive secret image sharing scheme with flexible-size shadows is proposed. First, a secret image is measured by compressed sensing (CS). Then, the quantized measurement values are divided into n shadows using Shamir's (t,n)-threshold scheme. At the receiver side, the secret image can be reconstructed if any t of n shadows are obtained, but fewer than t shadows reveal no information. Due to the fact that CS's reconstruction quality is flexibly adaptive to the number of measurements, our scheme features flexible shadow size and obtains the property of real-time and progressive transmission as well as error resilient. Experimental results show that the proposed scheme achieves better performance in view of the reconstructed secret image.
Signal Processing: Image Communication;29(1):128–134