2019
Annousakis-Giannakopoulos, Konstantinos; Ampeliotis, Dimitrios; Skodras, Athanassios
Could DCT Reveal Photorealistic Images? Proceedings Article
In: 2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA), pp. 1-5, 2019.
Abstract | Links | BibTeX | Tags: Computer Graphics, DCT, Digital forensics, Discrete Cosine Transform, Fake Photos, Photorealism, Quantization
@inproceedings{Annousakis-Giannakopoulos2019,
title = {Could DCT Reveal Photorealistic Images?},
author = {Konstantinos Annousakis-Giannakopoulos and Dimitrios Ampeliotis and Athanassios Skodras},
doi = {10.1109/IISA.2019.8900752},
year = {2019},
date = {2019-07-01},
booktitle = {2019 10th International Conference on Information, Intelligence, Systems and Applications (IISA)},
pages = {1-5},
abstract = {With the development of computer graphics rendering software, it has become extremely difficult to distinguish whether an image is computer generated or a natural one. Therefore, it is really important to device robust methods for correctly classifying these two categories of images. In this work, a new approach to face the above problem is developed that is based upon the discrete cosine transform (DCT) of an image, in the YCbCr color space. The statistical features extracted, have been tested in suitable databases and the remarkable results indicate that the proposed model has a great potential to be used in digital images forensics.},
keywords = {Computer Graphics, DCT, Digital forensics, Discrete Cosine Transform, Fake Photos, Photorealism, Quantization},
pubstate = {published},
tppubtype = {inproceedings}
}
With the development of computer graphics rendering software, it has become extremely difficult to distinguish whether an image is computer generated or a natural one. Therefore, it is really important to device robust methods for correctly classifying these two categories of images. In this work, a new approach to face the above problem is developed that is based upon the discrete cosine transform (DCT) of an image, in the YCbCr color space. The statistical features extracted, have been tested in suitable databases and the remarkable results indicate that the proposed model has a great potential to be used in digital images forensics.