Detecting blurred image splicing using blur type inconsistency. (2017)

• Record Type: Journal Article
• Authors: Zeng, Feng; Wang, Wei; Chen, Junjie; Tang, Min
• Is Part Of: International journal of innovative computing and applications. Volume 8:Number 1(2017)
• Page Start: 31
• Page End: 40
• Publication Date: 2017
• Subjects: blurred images -- image splicing -- splicing detection -- partial blur type -- blur estimation -- blind image restoration -- blur type inconsistency -- cepstrum-based image transformation -- image segmentation -- simulation -- forgery detection -- small spliced regions -- ringing effect; Evolutionary computation -- Periodicals; Neural networks (Computer science) -- Periodicals; Genetic programming (Computer science) -- Periodicals; Biologically-inspired computing -- Periodicals; Swarm intelligence -- Periodicals; Quantum computers -- Periodicals; 006.3
• Languages: English
• ISSNs: 1751-648X

Abstract

In a tampered blurred image generated by splicing, the spliced region and the original image may have different blur types. Splicing detection in this image is a challenging problem. In recent years, researchers have proposed various methods for detecting such splicing. In this paper, we propose a novel framework for image splicing detection based on partial blur type inconsistency. In this framework, after the cepstrum-based image transforming, a blur type classification parameter is extracted from the spectrum characteristics of spliced blurred image. The blurred image is restored based on the blur kernel which is constructed by estimating the blur parameters. Finally, a fine measure method is applied to segmentation inconsistent region in restored images that contain large amounts of ringing effect. Simulation results show the proposed method effectiveness in detecting forgery part in spliced images with different blur types. The proposed method has good robustness against lossy JPEG compression and noising, which outperforms the state-of-the-art methods for small spliced regions.