Detection and Localization of Region-Level Video Forgery with Spatio-Temporal Coherence Analysis

                                                                                                                  Cheng-Shain Lin and Jyh-Jong Tsay


     We conduct experiments over several video sequences manipulated temporal copy-and-paste [6] and exemplar-based texture synthesis [1] inpainting methods. Two data sets are used to test our approach. The first data set contains 11 test video sequences, which is numbered 1-11, are captured using a Canon IXUS 750 digital video camera. The second data set contains 7 test video sequences, which is numbered 12-18, are obtained from the PETS 2009 [38].

 

 
 Video Number
 
 Detection results for temporal copy-and-paste inpainting
       
     Detection results for texture synthesis inpainting

Description
 



             1
   Frame size:
       
320×240
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video.
(d) Detection results.


 



             2
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.

 



             3
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video.
(d) Detection results.
 

 



             4
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence.
(b) Removed regions. 
(c) Inpainted video.
(d) Detection results.
  



             5
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



             6
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



             7
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence with camera motion
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



             8
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence with camera motion
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



             9
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence with camera motion
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           10
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence with camera motion
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           11
   Frame size: 
      320×240
  a   b
  c   d
(a) Original video sequence with camera motion
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           12
  Frame size: 
      768×576
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
   



           13
   Frame size: 
      768×576
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



          14
  Frame size: 
      768×576
  a   b
  c   d
(a) Original video sequence
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           15
   Frame size: 
       720×576
  a   b
  c   d
(a) Original video sequence with multiple objects
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           16
   Frame size: 
       720×576
  a   b
  c   d
(a) Original video sequence with multiple objects
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           17
   Frame size: 
       720×576
  a   b
  c   d
(a) Original video sequence with multiple objects
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.
  



           18
   Frame size: 
       720×576
  a   b
  c   d
(a) Original video sequence with multiple objects
(b) Removed regions.
(c) Inpainted video. 
(d) Detection results.


[Matlab]
1. Sample Code: [download]
2. Test video sequences: [download]



REFERENCES
    [ 1 ]  A. Criminisi, P. Perez, and K. Toyama, “Region filling and object removal by exemplar-based image inpainting,” IEEE Trans. on Image Processing, vol. 13, no. 9, pp. 1200-1212, Sept. 2004.
    [ 6 ]  K. A. Patwardhan, G. Sapiro, and M. Bertalmio, “Video inpainting under constrained camera motion,” IEEE Trans. on Image Processing, vol. 16, no. 2, pp. 545-553, Feb. 2007.
    [38]  Pets 2009: Performance evaluation of tracking systems 2009 dataset, http://www.pets2009.net