Indian Journal of Science and Technology

Abstract

Background/Objectives: Copy-move is one of the familiar and crucialimage tampering techniques. To develop a method that automatically detect and locate the copy moved region is a primary objective. Methods/Statistical analysis: A novel method that employs Discrete Cosine Transform (DCT) to transform image spatial co-ordinates into frequency co-efficientis introduced. The 2D-DCT is applied to each sub blocks of tampered image and the high frequency co-efficient of each subblocks are extracted as a feature vector. These feature vectors are matched and tampered regions are located. Findings: The proposed method effectively detects and locates the tampered region in an image without using any authentication code or signatures. It also detects very small and multiple copy-move region. The efficiency of proposed method is evaluated on several test images. The results indicate that the accuracy of the method is high, number of false matches and dimension of feature vectors are reduced over the existing methods. Application/Improvements: The proposed method can be applied to criminal investigation, journalistic photography, law-enforcement,and medical imaging. The dimension reduction techniques may be implemented to reduce computation complexity and improve accuracy.

Keywords: Copy-move, Dimension Reduction, False Matches, Feature Vector, Tampering