Abstract: An authentication based on Shamir secret sharing with tampering detection and data recovery capability for document images is proposed, which includes two processes: tampering protection generation, and tampering detection and recovery. In the process of tampering protection generation, in order to improve the tampering detection accuracy and recover effect of existing document images authentication methods, a method of tampering detection signal generation based on weight is proposed in this paper. Firstly, it extracts characteristics for each 2×3 non-overlapping block. After that, the block characteristics and block content are put into the Shamir secret sharing scheme to generate multiple shares for tampered blocks detection and repair. Finally, the shares of the block constitute the α channel plane and combine with the original image to form a PNG(portable network graphic format)image. During the authentication process, the authentication signal extracted from the shares in the α channel plane can determine whether the image block has been tampered with, and can repair the content of the tampered image block through the reverse operation of Shamir secret sharing scheme. Experiments show that the proposed method not only has good tamper detection and the effectiveness of data repair, but also has good robustness against image clipping and noise attack.