Copy-move image forgery detection based on Gabor magnitude

Region Duplication Detection Based On Statistical Features of Image

“…The comparison is made on the basis of True detection accuracy (Ṫ) and false detection rate (Ḟ). Figure 8, Figure 9 and Figure 10 gives performance comparison for the proposed techniques with the state of art techniques [8,19,25]. Figure 9 and Figure 10 indicate that the proposed method performs well in comparison to methods [8,19,25].…”

“…The block based approach relies on dividing image into blocks either overlapping or non-overlapping [8,[19][20][21][22][23][24][25]. Features are extracted from each block and compared against each other to locate matching blocks to identify copy-pasted regions.…”

“…A slight decrease in the number of features in the feature vector results in a considerable improvement in speed and reduction in computational complexity [25]. This limitation has been a motivation to explore a new feature extraction technique that reduces the number of features representing each block and is robust against post processing operations.…”

“…For a large number of blocks, finding similar blocks will be cumbersome if brute force search is used. Thus lexicographical sorting of feature vectors is proposed by many authors [8, 19, 20, Copyright © 2017 SERSC 22,25]. Lexicographical sorting places similar feature vectors in adjacent rows.…”

“…The results obtained with the proposed method using GLRLM and GLCM are compared to the techniques based on DCT [8], PCA [19] and Gabor magnitude [25]. The comparison is made on the basis of True detection accuracy (Ṫ) and false detection rate (Ḟ).…”