Abstract: For robust image authentication,a robust hash scheme is developed for hierarchical authentication.The image's feature matrix in the wavelet frequency domain is extracted by dithering quantization.Then the feature matrix is divided into blocks.The level hash data are obtained by using the binary statistical quantity of all blocks and the authentication hash code is generated.For the robust and fragile features in the extraction,the authenticated hash can be robust to incidental manipulations and fragile to malicious tampering.Experiments are conducted to demonstrate the effectiveness of the hierarchical hash scheme for image authentication.