| Title |
Forensic Detection of Filtration Forgery in Digital Images |
| DOI |
https://doi.org/10.5573/ieie.2019.56.1.85 |
| Keywords |
영상포렌식 ; 블러검출 ; 필터커널 ; Dempster 조합법 |
| Abstract |
The main objective of image forensics is to detect correct and relevant symptoms that caused degradation of original images. In this paper, we proposed a new scheme to diagnose images by lowpass filtering that often used for compensating discontinuities in the forged images. Two different algorithms are adopted to measure the blur and estimate the filter kenel: edge based blur detection for filter size estimation (EBFE) mainly developed for NR image assessment and filter kernel detection by least energy based blind deconvolution algorithm (DCFE) mainly developed for image restoration. Final decision is made by Dempster’s combination rule. Results show 100% correct detection for the filter size less than 5 which is practically used, proving that the proposed scheme is effective for image Forensics. |