Abstract
A new approach to detecting forgery in digital photographs is suggested. The method does not necessitate adding data to the image (such as a Digital Watermark) nor require other images for comparison or training. The fundamental assumption in the presented approach is the notion that image features arising from the image acquisition process itself or due to the physical structure and characteristics of digital cameras, are inherent proof of authenticity and they are sensitive to image manipulation as well as being difficult to forge synthetically. Typically, such features do not affect image content nor quality and are often invisible to the inexperienced eye. The approach presented in this work is based on the effects introduced in the acquired image by the optical and sensing systems of the camera. Specifically, it exploits image artifacts that are due to chromatic aberrations as indicators for evaluating image authenticity.
Original language | English |
---|---|
Pages (from-to) | 71-91 |
Number of pages | 21 |
Journal | International Journal of Computer Vision |
Volume | 92 |
Issue number | 1 |
DOIs | |
State | Published - Mar 2011 |
Keywords
- Camera based forgery detection
- Chromatic aberration
- Image forgery
- Lateral chromatic aberration
- Lens artifacts
- Purple blooming
ASJC Scopus subject areas
- Software
- Computer Vision and Pattern Recognition
- Artificial Intelligence