A Quasi-Lossless Data Embedding Technique Based on Euclidean Distance Calculation

Recently data embedding over images have drawn tremendous interest, using either lossy or lossless techniques. Although lossy techniques can allow large hiding capacity, host image cannot be recovered with high fidelity. The aim of this work is to increase the embedding capacity for color palletized images while maintaining good visual quality of the stego image. The new technique is named "quasi-lossless " because the image histogram has been slightly changed but still satisfies high PSNR threshold that kept the stego image visually unchanged. The quasi-lossless technique is based on histogram analysis where least repeated colors are identified and manipulated to increase the embedding capacity. The proposed technique reported an embedding capacity that reached up to 60% and an improvement of more than 25% compared with other lossless method.