点击次数：

论文类型：期刊论文

发表时间：2009-07-01

发表刊物：CHINESE JOURNAL OF ELECTRONICS

收录刊物：SCIE、EI、Scopus

卷号：18

期号：3

页面范围：487-490

ISSN号：1022-4653

关键字：Digital image forensics; Photographic forgery; Color coherence; Copy-paste-blur tampering; Feature extraction

摘要：Exposing digital image forgeries is a major challenge for photography authentication and forensics investigation, which aims at proving the authenticity of digital photos. In this paper, we introduce a digital image forensics approach based on color coherence to exposing copy-paste-blur forgeries. We first discuss the inherent color coherence introduced by imaging pipeline, and then obtain several statistical features from the coherence characterization. Using the SVM classifier, we reveal traces of copy-paste-blur tampering in photographic forgeries. Experimental results indicate that the proposed method can effectively expose the copy-paste-blur forgeries and localize the tampered regions with high accuracy.