A logarithmic quantization index modulation for perceptually better data hiding

Area of Science:

• Digital Signal Processing
• Information Security
• Data Compression

Background:

• Quantization Index Modulation (QIM) is a watermarking technique.
• Logarithmic quantization offers perceptual advantages but previous methods had drawbacks.
• Mu-law companding is a standard for logarithmic compression.

Purpose of the Study:

• To propose a novel quantizer level arrangement for QIM using mu-law companding.
• To extend the method to vector quantization.
• To enhance watermark robustness and performance.

Main Methods:

• Utilizing the mu-law compression function to transform the host signal to a logarithmic domain.
• Uniformly quantizing the transformed data and inverting the transformation.
• Extending scalar quantization to vector quantization by quantizing magnitudes on hyperspheres with logarithmic radii.
• Calculating optimal mu parameter based on host signal distribution.
• Incorporating a secret key for enhanced security.

Main Results:

• The proposed method demonstrates stronger watermark embedding compared to conventional QIM.
• Performance analysis shows improved robustness and effectiveness.
• Simulations on artificial and real image data validate analytical derivations.
• The method overcomes limitations of previous logarithmic quantization algorithms.

Conclusions:

• The novel mu-law based logarithmic quantization for QIM provides superior watermarking performance.
• The scalar and vector quantization approaches are effective and robust.
• This method represents an advancement in digital watermarking techniques.