United-Judgment Methods Based on Parameter-Estimation for Image Steganalysis

In order to synthetically utilize the multi-steganalysis algorithms, and ulteriorly enhance the detection accuracy, the united-judgment methods based on parameter-estimation were proposed for image steganalysis. According to two types of universal blind detection and specific steganalysis, a united-judgment method based on weight and threshold, and also a method based on segment were proposed in this paper. Experiments were both made for the former method with seven kinds of typical blind detection algorithms and the latter one with five kinds of typical spatial domain steganalysis algorithms, respectively. The results showed that the proposed methods can synthetically utilize the existing detection algorithms, and comparing the results with a separate use of only one algorithm, the proposed methods could achieve higher detection accuracy.