A noise-based image splicing localization method for the case that the source images are with distinct ISO settings
Demo for reproducing the result of
[1] "Exposing Image splicing with inconsistent sensor noise levels" (accepted by Multimedia Tools and Applications, 2020 https://link.springer.com/article/10.1007/s11042-020-09280-z).
Homepage: http://www.escience.cn/people/Zenghui Contact: zengh5@mail2.sysu.edu.cn
This work is an improved version of our previous work
Zeng H, Zhan Y, Kang X, Lin X (2017), Image splicing localization using PCA-based noise level estimation. Multimedia Tools and Applications, 76(4):4783–4799
Run Splice_NL_ISO.m to see the the comparison results. The code has been verified on Matlab 2015a, Windows 10 64bit
In the caption of Fig. 1 in [1]
'... (left) A group of hikers from one ISO800 image is spliced into an ISO200 image.'
should be
'... (left) A group of hikers from one ISO800 image is spliced into an ISO100 image.'
In the caption of Fig. 5 in [1]
'... For display purpose, the noise levels in (a) and (b) are scaled to [0.4, 1.7], and the noise levels in (a) and (b) are scaled to [0.3, 2.3].'
should be
'... For display purpose, the noise levels in (a) and (b) are scaled to [0.4, 1.7], and the noise levels in (c) and (d) are scaled to [0.3, 2.3].'
We apologized for our carelessness
We public this code solely for the convenience of researchers to reproduce our idea accurately and rapidly. Hence, it is for academic use only.