印刷文件的數位鑑識

Abstract

由於眾多類型的犯罪調查和起訴的關係，近年來用來進行印刷文件源來設備的識別之數位鑑識技術愈顯重要。先前研究著重於透過掃描設備及其影像技術來取得精細的印刷文件之數位影像，然而這種技術受限於影像取樣的解析度與精緻程度不足，因此僅能提供有限的資訊量來進行相關的識別分析。相較之下，顯微技術可以取得這些印刷文件的高品質影像，呈現其文件內文中精緻印刷影像的形狀和表面紋理，如此即可以了解這些印表機列印文件時所呈現的詳細結構的特徵。 本研究主要目的在於探討經由掃描和顯微圖像技術應用於印刷來源設備識別的表現差異。目前已經有相關研究採取統計技術來辨識圖像特徵，例如：局部二進制模式（LBP）、灰度共生矩陣（GLCM）、離散小波變換（DWT）、空間濾波器、Wiener濾波器、Gabor濾波器、以及哈拉利克(Haralick)和分形特徵(fractal features)等。因此，這些影像資料經常被採取支持向量機（SVM）的分類機制做為資料分類識別使用。透過特徵濾波器的結合，加上決策融合計算技術，將可以提高文件辨識的準確率。 本研究的實驗結果顯示，採用所提出之掃描和顯微鏡圖像的分析技術，其平均辨識準確度預測值皆優於既存方法，因此可以證明本研究中所提出的決策融合模型和鑑識方法可以實際應用於現實世界做為印表機列印文件鑑識之技術使用。

Over the past several years, digital forensics for printed document source identification has begun to be increasingly important which can be related to the investigation and prosecution of many types of crimes. Prior researches implemented scanners as digitizing technique to resolve very fine printed document, but this technique provided limited information on the resolution and magnification of the sample. In contrast, the performance of microscopy techniques can retrieve the shape and surface texture of a printed document with detailed micro structures among printer sources. The primary purpose of this study is to investigate the scanned and microscopic images in printed source identification. There are many important statistical features implemented, i.e., Local Binary Pattern (LBP), Gray Level Co-occurrence Matrix (GLCM), Discrete Wavelet Transform (DWT), Spatial filters, Wiener filter, Gabor filter, Haralick, and fractal features. Consequently, they were utilized as separated sets of the features for classification using Support Vector Machines (SVM). The feature filter sets combination was also applied by the decision fusion calculation to achieve higher accuracy rate. The experimental results of this study indicate that the average accuracy prediction for scanned and microscopic images is superior to the existing approaches which demonstrate the decision fusion model proposed in this study can be promising for real world digital forensic applications.