Digital images can easily be tampered because of the popularity and power editing software. In order to create a persuasive forged image, the image is usually exposed to several geometric transformations, such as rescaling and rotating. Since the manipulations require a resampling step, uncovering traces of resampling became an important approach for detecting image forgeries. In this paper, we propose a new technique to reveal image resampling artifacts. The technique employs specific features of the linear dependencies of neighboring image samples for discriminating resampled images from original images. A machine learning method is utilized for classification. Experimental results in a large dataset show that the proposed technique is good in detecting resampled images, even when the manipulated images were slightly transformed.
CITATION STYLE
Nguyen, H. C. (2016). A machine learning based technique for detecting digital image resampling. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9622, pp. 75–84). Springer Verlag. https://doi.org/10.1007/978-3-662-49390-8_7
Mendeley helps you to discover research relevant for your work.