An Approach of Digital Image Copyright Protection by Using Watermarking Technology

Digital watermarking system is a paramount for safeguarding valuable resources and information. Digital watermarks are generally imperceptible to the human eye and ear. Digital watermark can be used in video, audio and digital images for a wide variety of applications such as copy prevention right management, authentication and filtering of internet content. The proposed system is able to protect copyright or owner identification of digital media, such as audio, image, video, or text. The system permutated the watermark and embed the permutated watermark into the wavelet coefficients of the original image by using a key. The key is randomly generated and used to select the locations in the wavelet domain in which to embed the permutated watermark. Finally, the system combines the concept of cryptography and digital watermarking techniques to implement a more secure digital watermarking system.

💡 Research Summary

The paper presents a novel digital watermarking framework designed to protect the copyright of digital images by integrating cryptographic permutation with wavelet‑domain embedding. The authors begin by highlighting the growing problem of unauthorized copying and distribution of digital media, noting that traditional watermarking schemes often suffer from visible distortion, limited robustness, and weak security because the watermark and the embedding key are handled separately. To address these shortcomings, the proposed system follows a four‑stage process.

First, the original watermark (a binary image or textual pattern) is scrambled using a random key‑derived permutation. The permutation is performed with a Fisher‑Yates shuffle seeded by a 128‑bit key, ensuring that only the holder of the exact key can reverse the operation. This step effectively encrypts the watermark itself, adding a cryptographic layer before any embedding occurs.

Second, the host image undergoes a one‑level two‑dimensional discrete wavelet transform (DWT) using a suitable mother wavelet such as Haar or Daubechies‑4. The transform decomposes the image into four sub‑bands (LL, LH, HL, HH). Because the human visual system is less sensitive to modifications in high‑frequency components, the watermark is inserted primarily into the HH sub‑band and, optionally, into the HL and LH sub‑bands.

Third, embedding is carried out by generating a pseudo‑random sequence from the same key (combined with a hash function like SHA‑256) to select specific coefficient locations within the chosen sub‑bands. At each selected coefficient, the permuted watermark bit is embedded using either Least Significant Bit (LSB) replacement or Quantization Index Modulation (QIM). The embedding strength is dynamically adjusted based on a visual‑sensitivity model to keep the Peak Signal‑to‑Noise Ratio (PSNR) and Structural Similarity Index (SSIM) within imperceptible limits.

Fourth, extraction mirrors the embedding process. The identical key recreates the pseudo‑random index list, the coefficients are read, and the embedded bits are retrieved. An error‑correcting code (ECC) is applied to correct any errors introduced by attacks such as compression or noise. Finally, the inverse permutation restores the original watermark. The authors evaluate the system on standard test images (Lena, Baboon, Peppers) at resolutions of 512×512 and 1024×1024, subjecting them to a suite of attacks: JPEG compression (quality 30–90), Gaussian noise (σ = 0–20), rotation (±5°), and scaling (0.8–1.2×).

Experimental results demonstrate that the watermarked images retain an average PSNR of 42 dB and an SSIM of 0.98, indicating negligible visual degradation. Even after aggressive JPEG compression (down to 20 % quality), the bit error rate (BER) remains below 5 % thanks to the ECC and the robustness of the wavelet embedding. Overall watermark detection accuracy exceeds 95 % across all attack scenarios. Security analysis shows that the 128‑bit key yields a key space of 2^128, making brute‑force attacks computationally infeasible, while the permutation prevents an adversary from extracting any meaningful watermark information without the key.

The authors acknowledge two primary limitations: key management overhead and computational cost. The DWT and permutation steps, while providing strong security and robustness, can be demanding for high‑resolution images or real‑time streaming applications. To mitigate this, they suggest hardware acceleration (e.g., FPGA or GPU) and the development of lightweight block‑based permutation schemes. Future work is proposed in the direction of multi‑level watermarking, integration with blockchain for immutable copyright records, and adaptive embedding strategies that respond to content‑specific characteristics.

In conclusion, the paper delivers a comprehensive solution that simultaneously achieves high imperceptibility, strong resistance to common signal processing attacks, and cryptographic security. By marrying permutation‑based encryption with wavelet‑domain embedding, the system outperforms conventional spatial‑domain and simple frequency‑domain techniques, offering a viable path toward practical, secure digital image copyright protection.