Scalable image tamper detection and recovery based on dual-rate source-channel coding

In image tamper detection and recovery methods prevalently two sets of data need to be embedded into the main image, one known as authentication data for tamper detection and the other known as reference data for tamper recovery. In this paper, a scalable method for image tamper detection and recovery based on a dual-rate source-channel coding is proposed. In the proposed method the original image is first compressed by a scalable source coding algorithm. The compressed bitstream is then partitioned into two parts. Both parts are protected with a channel coding algorithm but with two different rates according to their importance for tamper recovery process. The proposed method takes advantage of two state-of-the-art algorithms, SPIHT for source coding and LDPC for channel coding. Simulation results show a noticeable improvement compared with related tamper detection and recovery schemes in the literature. Besides the reconstruction quality of the recovered image is scalable. This means that in low tampering rates, high-quality images are recovered and in high tampering rates, the image is still recoverable but with less reconstruction quality. Therefore the proposed method succeeded to both recover higher tampering rates and preserve the image quality.