A DWT-based fragile watermarking tolerant of JPEG compression

In this paper, we propose a novel DWT-based fragile watermarking scheme by exploiting the characteristics of the human visual system (HVS). Multi-resolution tamper detection and image fusion are applied to watermark detection and accurate tamper region location. To distinguish between incidental and malicious modifications, some effective rules are presented. Mathematical morphology operations are used to improve the detection performance. Experiment results demonstrate that the watermark is fragile to malicious tamper and robust against high quality lossy image compression.