在電腦應用，網際網路無遠弗屆的時代，許多的媒體都以數位的形式記錄、呈現及傳播，數位設備的普及，相對地也面臨許多新興的數位需求，而數位鑑識即是其中之一。 鑑定相機來源之數位鑑識是研究數位相機與其所拍攝出之影像之關係，針對不同相機所拍攝出的影像，結合影像處理技術與支撐向量機SVM (support vector machine)方法，計算出影像的特徵值，並將這些特徵進行訓練及分類，用以鑑別該影像之來源相機；同時，依據分析結果提出特徵值分類訓練的改進方法以提高影像來源的鑑識率，實驗結果也說明其有效性。 相機來源之辨識方法係利用各家廠商問相機的內部成像演算法之不同，以影像特徵值的分類來歸納出隱藏在影像中的特徵，此研究除了在不同廠牌的相機間具有優異辨識率外，也探討在同廠牌且型號接近情況下，該方法是否仍具有辨識能力，並針對結果提出以影像特徵值來鑑識相機來源之可能限制。

In this digital era, all kinds of media data are recorded and presented digitally. The popularity of digital devices also brings up the needs for new digital data applications, and digital forensics is one of them. This study is to analyze the relationship between digital cameras and images produced by them. Digital image processing technology is first applied in the identification procedures in order to get images' features. Those features are later trained and classified to identify the images' original sources by using SVM (support vector machine) techniques. In addition, the identification features are further analyzed for improvements for enhancing the precision rate of the identification. Since camera manufacturers have various designs of image formation processes for digital cameras, the distinctive digital image feature will be identified and analyzed. This research not only identifies the camera sources among different manufactures, but also those of the same manufacturer with similar models. Therefore, the potential limits of this identification method will be verified by the research results.