A semi-fragile watermark algorithm based on the Walsh-Hadamard transform and convolution coding was proposed. In this algorithm

the watermark content of the host image was extracted by differential coding and convolution coding respectively based on the image block before embedding watermark

and was embedded on the energy value of the host image after Walsh-Hadamard transforming. In the case of watermark detection

the tampered place of the image was located by detecting the error bits generated by convolution decoding

and its result was used to recover the tampered image. The efficiency and the influence on the host image quality of the proposed algorithm were discussed theoretically and experimentally. The experimental results show that the proposed algorithm has high robustness against lossy compression

can detect and locate the place being maliciously tampered accurately and recover the lost image content roughly.