Optimization of Bit Plane Combination for Efficient Digital Image Watermarking

In view of the frequent multimedia data transfer authentication and protection of images has gained importance in todays world. In this paper we propose a new watermarking technique, based on bit plane, which enhances robustness and capacity of the watermark, as well as maintains transparency of the watermark and fidelity of the image. In the proposed technique, higher strength bit plane of digital signature watermark is embedded in to a significant bit plane of the original image. The combination of bit planes (image and watermark) selection is an important issue. Therefore, a mechanism is developed for appropriate bit plane selection. Ten different attacks are selected to test different alternatives. These attacks are given different weightings as appropriate to user requirement. A weighted correlation coefficient for retrieved watermark is estimated for each of the alternatives. Based on these estimated values optimal bit plane combination is identified for a given user requirement. The proposed method is found to be useful for authentication and to prove legal ownership. We observed better results by our proposed method in comparison with the previously reported work on pseudorandom watermark embedded in least significant bit (LSB) plane.

💡 Research Summary

The paper “Optimization of Bit Plane Combination for Efficient Digital Image Watermarking” proposes a novel spatial‑domain watermarking scheme that improves both robustness and transparency by carefully selecting which bit‑plane of the host image should carry the watermark. Unlike many earlier works that embed a pseudorandom pattern into the least‑significant bit (LSB) plane, the authors use a digital‑signature watermark—an intentionally strong, uniquely identifiable pattern—so that the presence of the watermark can serve as legal proof of ownership.

The core idea is to replace a “significant” bit‑plane of the original gray‑scale image with a high‑strength bit‑plane of the signature watermark. After extensive testing, the authors found that the 7th bit‑plane (the bit just above the LSB) of the host image provides a good trade‑off: it is still visually imperceptible to the human visual system (HVS) while being considerably more resistant to common attacks than the LSB. The watermark’s 1st bit‑plane (the most significant bit of the signature) is embedded into this host plane.

To determine the optimal combination of host‑image and watermark bit‑planes, the authors define a systematic evaluation framework. Ten representative attacks are selected: rotation, affine transformation, quantization, 41 % cropping, low‑pass filtering, translation, salt‑and‑pepper noise, contrast stretching, JPEG compression, and scaling down. For each attack, the correlation coefficient (CRC) between the extracted watermark and the original signature is computed; CRC values range from 0 (no similarity) to 1 (perfect recovery). Recognizing that different applications may prioritize different attacks, the authors introduce user‑defined weightings (a_i) for each attack, constrained such that (\sum_{i=1}^{10} a_i = 1). A weighted correlation coefficient (WCRC) is then calculated as

\