Efficient gray-level digital image watermarking based on vector quantization

Digital watermarking has been presented as a new method for copyright protection by embedding a secret signal in a digital image or video sequence. Common digital image watermarking techniques are based on the concept of spread spectrum communications, which can be classified in two catalogues: spatial domain and transform domain based. Most of transform domain watermarking methods are based on discrete cosine transforms (DCT) and robust to JPEG lossy compression. Recently, digital image watermarking based on another important lossy compression technique, vector quantization (VQ), has been presented, which carries watermark information by codeword indices. It is secret and efficient, and is robust to VQ compression with the same codebook. However, the embedded information is less and the extraction process requires the original image. This paper presents a more efficient VQ based image watermarking method, which can embed a large gray level watermark into the original image with less extra distortion and perform the watermark extraction without the original image. In addition, the proposed watermarking algorithm is very secret because two keys are required for watermark extraction. Experimental results demonstrate the effectiveness of the proposed technique.