An Efficient Vector Quantization Based Watermarking Method for Image Integrity Authentication

This paper presents a two-stage watermarking technique for image authentication adapting advantages of vector quantization (VQ). In the present algorithm robust watermark and semifragile watermark are embedded independently in VQ compressed image in two successive stages. Robust watermark and VQ enhances the security of the system by providing double protection to designed system. A quantitative threshold approach using pixel surrounding error pixel is suggested for identification of attacks as acceptable or malicious. Experimental results demonstrate the capabilities of the method in classifying attacks and correctly locating tamper location. It is possible to detect and determine tamper with very high sensitivity. Present scheme outperforms previous algorithms and can distinguish malicious tampering from acceptable changes, and tampered regions are localized accurately.