@inproceedings{bc2522474d6e4196b56cdf917bc61fde,
title = "A digital watermarking method based on classified labeled-bisecting-k-means clustering",
abstract = "With the growing interest in copyright protection of multimedia, digital watermarking has been presented and widely researched. The digital image watermarking techniques are usually performed in spatial or transform domain. Most of transform-domain watermarking methods are based on Discrete Cosine Transforms (DCT) and robust to JPEG lossy compression. Recently, digital image watermarking based on another significant lossy compression technique, Vector Quantization (VQ), has appeared, which carries watermark information by codeword indices. It is quite effective, especially for the robustness to VQ compression with the same codebook. We develop a more favorable image watermarking method based on classified VQ in this paper. This technique can achieve excellent embedding performance and extract the watermark without the original image. This original theoretical result is acceptable with experiments that show superiority to some existing VQ-based watermarking algorithms, and the watermarked image is robust to some common signal processing operations.",
keywords = "Classified Vector Quantization, Data Hiding, Digital Watermarking, Image Processing, Vector Quantization",
author = "Wen Xing and Lu, \{Zhe Ming\} and Wang, \{Hao Xian\}",
year = "2003",
language = "英语",
isbn = "0780378652",
series = "International Conference on Machine Learning and Cybernetics",
pages = "2891--2895",
booktitle = "International Conference on Machine Learning and Cybernetics",
note = "2003 International Conference on Machine Learning and Cybernetics ; Conference date: 02-11-2003 Through 05-11-2003",
}