Affiliations: [a] Department of Computer Science and Engineering, Indian Institute of Technology (Indian School of Mines), Dhanbad, India
| [b] Department of Computer Science and Engineering, National Institute of Technology, Delhi, India
| [c] Division of Information Technology, Netaji Subhas Institute of Technology, Delhi, India
Correspondence:
[*]
Corresponding author. Choudhary Shyam Prakash, Department of Computer Science and Engineering Indian Institute ofTechnology (Indian School of Mines), Dhanbad, India. E-mail: [email protected].
Abstract: Copy-move forgery is one of the famous manipulation technique in digital image. Many block-based techniques have been proposed previously for forgery detection, but most of them have higher computational complexity due to higher number of feature vectors dimension. In this paper, we have tried to reduce the feature vectors dimension. This paper proposes a copy-move forgery detection (CMFD) technique based on circular blocks and discrete cosine transform (DCT) with fewer feature vectors than the prevalent methods. Initially an input image is taken and divided into overlapping blocks. To extract the features from each block, DCT transformation is used on each block. Then, these features are represented using a circle block to reduce the feature vectors dimension. The extracted feature vectors are then used for matching process to locate the manipulated regions. Experimental results depict the performance of the proposed method and robustness against the post-processing operations. The computational complexity of the proposed method is lower than the existing techniques due to fewer feature vectors dimension.
