Software

Nonparametric Steganalysis of QIM Steganography Using Approximate Entropy

Free registration required

Executive Summary

This paper proposes an active steganalysis method for Quantization Index Modulation (QIM) based steganography. The proposed nonparametric steganalysis method uses irregularity (or randomness) in the test-image to distinguish between the cover-image and the stego-image. The authors have shown that plain quantization (quantization without message embedding) induces regularity in the resulting quantized-object, whereas message embedding using QIM increases irregularity in the resulting QIM-stego. Approximate entropy, an algorithmic entropy measure, is used to quantify irregularity in the test-image. The QIM-stego image is then analyzed to estimate secret message length. To this end, the QIM codebook is estimated from the QIM-stego image using first-order statistics of the image coefficients in the embedding domain.

  • Format: PDF
  • Size: 992.24 KB