Improving RLRN Image Splicing Detection with the Use of PCA and Kernel PCA

Table 4

Comparison between proposed approaches and other methods.

Feature extraction methods	Dimensionality	TP (%)	TN (%)	Acc. (%)
Expanded DCT Markov [10]	100	89.92	90.21	90.07
DWT Markov [10]	100	87.58	85.39	86.50
Expanded DCT Markov + DWT Markov [10]	100	93.28	93.83	93.55
HHT + moments of characteristic functionswith wavelet decomposition [5]	110	80.25	80.03	80.15
Run length and edge statistics based model [13]	163	83.23	85.53	84.36
RLRN + kernel PCA (proposed)	50	90.38	86.18	88.28