Impact of watermarking on offline signature verification in intelligent bio-watermarking systems

Bio-watermarking systems were introduced as the synergistic integration of biometrics and digital watermarking to assure the integrity, authenticity and confidentiality of digitized image documents, and biometric templates. In this paper, the impact of watermarking attacks on the performance of offline signature verification is assessed in the context of intelligent bio-watermarking systems. The considered system is based on incremental learning computational intelligence, and multi-objective formulation that allows optimizing parameters according to watermark quality and robustness simultaneously. In this study, Extended Shadow Code features are extracted from digitized offline signatures, collected into feature vectors, and discretized into binary watermarks prior to being embedded into high resolution grayscale face image. The impact on biometric verification performance of quantization and different intensities of attacks are considered, and also observed the impact of using only certain areas of face images of higher texture Region Of Interest (ROI) for embedding the watermark. Experimental results conclude the optimal discretization, and better watermark fitness and verification performance when embedding in ROI. To improve the performance in future research, the authors propose to embed more reference signatures, use efficient ROI identification techniques, and finally novel formulation to add biometrics verification fitness to the watermark quality and robustness fitness during embedding optimization. The proposed system can be applied for verifying individuals crossing borders using offline signatures, or protecting biometric templates.