Dual Water Marking Based Image Collection in Multimedia Sensor Networks

This study presents a novel image collection scheme for multimedia sensor networks based on dual water marking. The dual water marking includes two folds: one is robust water mark for the identification of the sender and freshness authentication of images and the other is the fragile water mark for the tamper detection and recovery of image contents. In this scheme, the sensor node groups image frames and two successive frames compose the non-overlapping authentication and restoration group. The dual water marking bits are computed from the first image and embedded into the image group. The sink performs the verification and recovery by dual water mark. Compared with previous work, our approach can not only implement the task of identification and authentication, but also improve image quality. Experimental results show that our scheme achieves significant gains in terms of identification and authentication performance and efficient packet loss tolerance to improve the image quality.