提出通過對(duì)分塊圖像的DCT 系數(shù)進(jìn)行動(dòng)態(tài)范圍壓縮來改進(jìn)傳統(tǒng)的基于DCT 變換的圖像自嵌入水印算法，并結(jié)合灰度變換函數(shù)與JPEG 標(biāo)準(zhǔn)量化表重新設(shè)計(jì)了DCT 系數(shù)碼長(zhǎng)分配表，大幅度提升了量化過程保留的圖像頻段范圍，使水印圖像在遭受篡改攻擊后能較好地恢復(fù)；最后，給出了MATLAB 的算法實(shí)現(xiàn)。
關(guān)鍵詞：DCT；水印；MATLAB；碼長(zhǎng)；動(dòng)態(tài)壓縮；自嵌入
Abstract: In this paper, a better scheme of self2embedding an image into itself is proposed by compressing the dynamic range of DCT coefficients , whose bit lengths are nicely chosen according to gray transforming function and standard JPEG quantization table. By this scheme , the image’s frequency segment is largely extended which could be saved, and the watermarking image tampered could be properly recovered . In addition ,the key algorithm,
which comes true in MATLAB, is amply analyzed in this paper.
Key words: DCT ; watermark; MATLAB; code length; dynamic compression; self2embedding