For the problem that the prediction accuracy of the traditional median edge predictor al-gorithm is relatively low in the absence of edges,a high-capacity medical fragile reversible water-marking algorithm based on 8-neighborhood edge prediction is proposed.With the consideration that the influence of the 8 pixels in the neighborhood of the prediction point varies in different direc-tions,it calculates and optimizes the weights in the corresponding positions,and compares the pre-diction values of the 8-neighborhood with those of the median edge,so as to obtain a more accurate prediction.Meanwhile,to avoid the waste of restoring pixel points,the overall pixel points are di-vided into regions,and the method of multi-layer embedding is adopted for the prediction point pix-els,which significantly increases the embedding capacity.Finally,the watermark information is embedded into the carrier image by improving the method of the predicted error histogram shifting.Experiments show that the algorithm can achieve the reversible recovery of medical images,and al-so meet the needs of large-capacity embedding.At the same time,it has a strong vulnerability to most attacks,and can protect the security of watermark images.
维普
万方数据