Effectiveness of LSB and MLSB information embedding for BMP images

Abstract

The paper presents a study of efficiency of information embedding in BMP images by using codes perfect in a weighted Hamming metric (WHM) and standard Hamming codes. The weighted container model and modification of algorithm F5 is used, which allows to record information in multiple bit planes images (MLSB), taking into account the significance of each bit plane. In this paper the effectiveness of such weighted embedding approach is compared with the common matrix information embedding method that uses classic Hamming code (7,4,3). The effectiveness of information embedding is estimated by PSNR metric and by special Penalty function value. Also composition of RGB color channels in different order is attempted to define the best weighted container structure taking into account Human Vision System features.

Keywords

• Steganography
• Error-correcting codes
• weighted Hamming metric
• F5
• Weighted F5
• Penalty Function
• Weighted container optimization
• MLSB embedding