| We propose a secure steganographic approach with a high capacity. We first use small neighborhoods of the original histogram to calculate the minimum remapping from a histogram to the odd-only and even-only histograms. We then employ a quantization index modulation-based odd-even embedding to preserve the histogram of each embedding subchannel. A minimum distortion remapping policy is used to decide the quantization value so the coefficient is increased or decreased according to a constant probability. Our embedding ensures the modification occurs in a small neighborhood and the details of the original histogram are maintained. We extensively compare our approach with two state-of-the-art steganographic approaches, MB1 and MB2, using Fridrich's 23-feature-based blind steganalysis tool with a linear discriminant analysis. The results show that our method achieves better image quality, better embedding capacity, and comparable security to MB I and MB2. |
