Open AccessImage Forgery Detection with SIFT and RANSAC
Author(s) -
Satish B Pratapur*,
Dr.Shubhangi D.C
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.c4245.098319
Subject(s) - ransac , scale invariant feature transform , artificial intelligence , similarity (geometry) , image (mathematics) , pattern recognition (psychology) , computer science , computer vision , transformation (genetics) , rotation (mathematics) , scaling , multidimensional scaling , mathematics , machine learning , biochemistry , chemistry , geometry , gene
in this paper, simulations were performed using SIFT and RANSAC to highlight the forged regions in the doctored image. SIFT algorithm is modified to consider unit vectors as the features of the blocks. Blocks with similar unit vectors were grouped into cluster. Mean values of the clusters were compared to determine the similarity between clusters. Once the clusters were formed, the image was subjected to RANSAC algorithm to determine the geometric transformation and to highlight the forged region in the doctored image. Two simulations were performed to test the performance of the proposed method. First, doctored image with only scaling and next, image with both scaling and rotation were tested. The simulation results are presented in detail.