可應用各種強韌及高位元率之編碼方式的圖形式軟體浮水印架構

Abstract

在這篇論文中，我們提出一套圖形式軟體浮水印架構，可以依據不同的安全需求，適度地採用適用的浮水印編碼演算法，以兼顧運作效能與安全度。現有的圖形式軟體浮水印演算法中，如QP及QPS，僅能用於解決特定問題，因此無法提供一套全面性的解決方案，符合軟體浮水印在效能與安全方面的各種需求。此外，這些演算法具有低資料率與遭受惡意攻擊而破壞浮水印資訊等缺點。為了解決現有浮水印編碼演算法中的缺點，在我們所提出的軟體浮水印架構中，可以視執行效能與安全需求，利用提出的三種編碼演算法之一，將浮水印資訊嵌入於軟體模組中，以提供更高的資料率與強韌度。在這個浮水印的架構中，我們將會把程式轉換成圖形，並且將這個圖形分割成較為小的子圖，根據安全上或效能上的多樣需求，這些子圖將可以依照不同的編碼機制嵌入浮水印資訊。最後，我們分析並比較各種現存的軟體浮水印機制與本論文所提之方法，在錯誤偵測、抵擋惡性攻擊以及資料率等能力，以評估其強韌度與效能。

In this paper, we propose a graph-based software watermarking framework to flexibly adapt hybrid encoding algorithms according to different security requirements. Existing graph-based software watermarking algorithms, such as QP and QPS, only address specific problems and thus cannot provide a one-fit-all solution to meet various requirements in terms of performance and security. In addition, these algorithms could suffer from the low data rate issue and vulnerable to additive and subtract attacks. To address the above issues, the proposed framework works in a hybrid manner and three encoding algorithms are also proposed to cooperate with our framework and to achieve higher data rate and robustness. In this paper, a software program is represented as a graph and can be further divided into smaller sub-graphs. The watermarking procedure runs through the graph and applies one of the three proposed encoding algorithms to each visited sub-graph per the security and efficiency requirement. As an evaluation of our work, error detection capability, attack resistance and encoding data rate are analyzed and compared between our work and the related work. The result shows that the proposed framework performs better bit rates under the same requirements.