應用Toral Automorphism並結合機率理論之易碎型浮水印研究

Abstract

科技發展讓影像數位化成為可能，但也伴隨著版權侵犯等問題的出現。本研究專注於用於版權保護的數位浮水印，目標是改善視覺系統難以察覺的影像修改與偽造問題，加以抵抗各種攻擊，如:大範圍攻擊、雜點攻擊。本研究使用易碎浮水印認證技術，搭配Toral Automorphism之應用，透過最低有效位元置換法（LSB）將版權資訊嵌入影像。然而，LSB面臨著資訊量與影像破壞、偵測錯誤率之間的平衡問題。本研究核心在於尋找這個平衡，以及建立影像區塊之間的連結。實驗結果表明，本論文的方法在抵抗攻擊及正確標示率有些微提升。

The development of technology has made digitalization of images possible, but it has also brought about issues such as copyright infringement. The goal is to im-prove the issues of imperceptible image modification and forgery in the visual system and to resist various attacks, such as large-scale attacks and random noise attacks. This research utilizes fragile watermark authentication technology, combined with the application of Toral Automorphism, embedding copyright information into images using the Least Significant Bit (LSB) substitution method. However, LSB faces a trade-off between the amount of information and issues like image degradation and detection error rates. The core of our research lies in finding this balance and establishing connections between image blocks. Experimental results demonstrate that our approach performs remarkably well in resisting attacks and providing accurate labeling.