The capacity estimation approach for digital watermark in JPEG images

The watermarking quality is determined by three factors such as robustness, imperceptibility and capacity, in which capacity is the most important. When capacity is high, the robustness of watermarking is increasing. However the imperceptibility is decreasing, so it is worthwhile to ensure the accuracy of data hiding. In this paper we propose a new approach to estimating the data hiding capacity of JPEG-compressed images, which not only takes the human visual system into consideration, but also adopts adaptive quantization table. We test the, capacity in terms of different JPEG compression by using the approach, and compare it with Waston's model. The experimental results show that the capacity has greatly increased under certain watermark robustness and security.