SECURE TEXT HIDING IN AUDIO STEGANOGRAPHY USING SHUFFLED ARNOLD CAT MAP BASED ENCRYPTION AND ADAPTIVE STEP SIZE QUANTIZATION INDEX MODULATION
Keywords:
Adaptive Quantization Index Modulation, Audio steganography, Embedding, Hiding text, Shuffled Arnold Cat Map.Abstract
Audio steganography offers promising solution by enabling embedding of secret messages in audio files. However, achieving high security, robustness against attacks, and minimal distortion of host audio signals remains significant challenge. This research addresses challenges of imperceptibility in hiding sensitive frequency regions of messages ensuring robustness against audio processing and encrypts message for added security. Shuffled Arnold Cat Map (SACM) and Adaptive Step Size Quantization Index Modulation (ASSQIM) are proposed to encrypt secret data and hide text in audio steganography. ASSQIM embeds secret messages by adjusting step size based on frequency analysis. Lower step size is used in sensitive frequency region to manage imperceptibility, whereas larger step size is applied in nonsensitive regions to enhance robustness against attacks. Hidden message is encrypted using SACM from cover audio, which enhances security under different attacks. Experimental analysis shows SACM-ASSQIM obtains high Normalized Correlation (NC) of 0.9999 compared to QIM with reduction in Log Spectral Distance (LSD) from 0.998 to 0.944 representing significance on enhanced robustness and security with audio-text. Additionally, SACM- ASSQIM obtains high Peak-to-Signal-Noise Ratio (PSNR) of 29.4726 dB in high-pass filter attack and 98.3547 dB on TIMIT and GTZAN datasets when determining with audio-image compared to Linear Predictive Analysis (LPA).
