為了減少儲存空間的浪費並加快傳輸速度,大多數的數位影像都會經過失真壓縮處 理。由於失真壓縮會使得資料損失,因此,使用易碎型浮水印技術會導致失真壓縮格 式的影像無法通過驗證。相對於易碎型浮水印對於攻擊的敏感性,半易碎型浮水印則 可以抵抗失真壓縮與某些常見的影像處理,並對於惡意的竄改保有易碎性。2011 年 Wenyin and Shih 學者利用區域二元特徵(local binary pattern, LBP)運算子提出一套半易 碎型浮水印方法,但此方法存在著安全性方面的問題,有可能無法偵測出攻擊者的惡 意竄改;此外,該方法嵌入浮水印的規則,在某些情況下可能會導致較差的影像品質。 因此,本研究將提出一套改善的方法,可解決 Wenyin and Shih 等學者方法的安全性問 題,並且結合整數線性規劃的技術,來提昇嵌入浮水印後的影像品質。 Fragile watermarking technique is used for image authentication. However, some common used lossy compression cannot pass the tampering detection even if that is used to reduce the image size hence cannot be seen as an malicious attack. Therefore, some researchers proposed semi-fragile watermarking schemes, which do not recognized lossy compression as an malicious attack. In 2011, Wenyin and Shih proposed a semi-fragile watermarking scheme based on local binary pattern (LBP) operators, but the watermarked image can be tampered with without being detected by their scheme. Besides, the watermarked image quality may be degraded under some circumstances. Therefore, this research will propose an semi-fragile watermarking scheme to improve the security of Wenyin and Shih’s scheme. In addition, integer linear programming is adopted in our scheme to upgrade the watermarked image quality.