在各種犯罪活動(dòng)中文檔偽造是一種常見的偽裝手段,由于其易于實(shí)施而難于分辨,文檔偽造的檢測技術(shù)顯得更加重要而迫切,通過對文檔中筆跡改動(dòng)的檢測辨別文檔真?zhèn)问瞧渲兄匾氖侄�。然�?篡改者往往使用相近顏色的墨水對文檔進(jìn)行篡改,因此往往無法靠眼睛進(jìn)行分辨。數(shù)字圖像處理的方法提供更加精細(xì)的分辨能力,因此為這一問題帶來新的解決方案。本文針對文檔篡改中的筆跡檢測問題,提出了兩種基于圖像處理的檢測方法。論文的主要?jiǎng)?chuàng)新性工作包括:首先,綜述并分析了用于筆記檢測的重要圖像特征,對比了不同的色彩空間表達(dá)。鑒于基于數(shù)字圖像的筆跡檢測算法目前缺少測試樣本,我們首先建立了一個(gè)包含各種篡改情況的樣本庫,并通過掃描儀對樣本進(jìn)行數(shù)字化,得到了一個(gè)數(shù)字圖像樣本庫,為后續(xù)的算法研究提供支撐;第二,提出了一種基于RGB特征距離的篡改筆跡檢測方法,通過提取RGB通道的9個(gè)特征并計(jì)算特征與文檔中所有目標(biāo)的距離進(jìn)行檢測。首先文檔被分割成獨(dú)立的目標(biāo)對象,然后計(jì)算每個(gè)目標(biāo)對象的9個(gè)特征,以及用均方根誤差表示的兩兩特征向量之間的距離。然后通過改進(jìn)的Thompson Tau算法檢測異常點(diǎn)并進(jìn)行第二輪異常點(diǎn)檢測。實(shí)驗(yàn)證明了所提出方法在檢測和定位... 

【文章來源】：哈爾濱工業(yè)大學(xué)黑龍江省 211工程院校 985工程院校

【文章頁數(shù)】：89 頁

【學(xué)位級別】：碩士

【文章目錄】：
摘要
ABSTRACT
Chapter 1 General Introduction
    1.1 What is the handwriting forgery
    1.2 Background and history of handwriting forgery
        1.2.1 Famous forgery cases of the 20th century
    1.3 Features of handwriting
        1.3.1 Characteristics of handwriting
    1.4 Types of handwriting forgery
    1.5 Research abroad and limitations
        1.5.1 Analysis of literatures
    1.6 Problem statement and motivation
    1.7 Thesis outline and contributions
Chapter 2 Related Works
    2.1 Color features
        2.1.1 Color models
        2.1.2 Color histogram
    2.2 English handwriting dataset for forgery detection
        2.2.1 Dataset description
    2.3 Summary
Chapter 3 Handwriting Forgery Detection Based On Ink Color Features
    3.1 Introduction
    3.2 Proposed method
        3.2.1 Pre-processing and segmentation
        3.2.2 Features extraction
        3.2.3 Distance measurement
        3.2.4 Outliers detection
    3.3 Experiment results
        3.3.1 Detection performance
        3.3.2 Experiment results and discussion
    3.4 Summary
Chapter 4 Handwriting Forgery Detection Based On Structural Similarity Index of Color Histogram
    4.1 Introduction
    4.2 Proposed method
        4.2.1 Pre-processing and segmentation
        4.2.2 Color histogram
        4.2.3 The structural similarity index measure
        4.2.4 Similarity matching and forged object detection
    4.3 Experiment results
        4.3.1 Visual results
        4.3.2 Detection performance
        4.3.3 Running time
    4.4 Summary
Conclusions
References
List of Publications
Acknowledgements



本文編號：3512730