離散小波轉換應用於彩色雷射印表機鑑識之研究

Abstract

隨著個人列印輸出設備的普及，利用相關設備從事的偽造犯罪事件愈趨頻繁。本研究針對前述犯罪辨識的需求，提出以已知的印刷影像辨識印表機的方法。首先，將各印表機列印的影像結果經掃描後，透過小波轉換分解出LH、HL、HH 3個擁有高頻訊號的子頻帶，採取常見的統計方法獲得45個影像特徵值，並利用最佳化特徵選擇演算法選出較佳之特徵後，進一步使用SVM 建立分類模組去辨識未知的印刷影像。實驗結果顯示，我們所提出之方法在彩色雷射印表機具有良好之辨識效果。

With the popularity of individual printer devices, the use of these related print device frequently increases criminal counterfeiting activities. Based on the needs of identifying crime activities, this study offers a method to identify color laser printer with printed color images. First, the printed color image is scanned. Then 45 statistical features are extracted from LH, HL, HH 3 sub-bands that have high frequency coefficients after the wavelet transform. Next step, the best features are selected, base on SVM classification module so we can identify the source of unknown printed image. Experimental results demonstrate that the proposed method can effectively identify the correct color laser printer with high accuracy.