Detection of colorization based image forgeries using convolutional autoencoder method
Soumyashree Muralidhar Panchal, Asha Kethaganahalli Hanumanthaiah, Bindushree Channabasavaraju Doddasiddavanahalli, Manju More Eshwar Rao, Ambika Belekere Jayaramu
Abstract
Recently, it has become difficult to recognize and easier to misuse digital images due to the large number of editing tools available. Detecting forgeries in images is crucial for security and forensic purposes. Therefore, this research implements a deep learning (DL) method of convolutional autoencoder (CAE) which improves colorization-based image forgery detection by leveraging spatial and color information, increasing the detection accuracy. At first, the pre-processed input forgery images are used with the wiener filtering-contrast restricted improved histogram equalization (WE-CLAHE) technique. Hybrid dual-tree complex wavelet trigonometric transform (H‑DTCWT) and VGG-16 are used to extract effective features from the clustered data. Improved horse herd optimization (IHH) is employed to reduce the dimensionality of a feature. At last, the CAE model is implemented to significantly recognize the image forgery. The accuracy of CASIA V1 and GRIP datasets of 99.95% and 99.97%, respectively is achieved. Hence, this implemented method obtains a high forgery detection performance than the existing methods.
Keywords
Blind forgery detection; Convolutional autoencoder; Deep learning; Digital images; Image forgery
DOI:
http://doi.org/10.11591/ijeecs.v36.i2.pp1114-1126
Refbacks
- There are currently no refbacks.
![Creative Commons License](https://i.creativecommons.org/l/by-sa/4.0/88x31.png)
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Indonesian Journal of Electrical Engineering and Computer Science (IJEECS)
p-ISSN: 2502-4752, e-ISSN: 2502-4760
This journal is published by the Institute of Advanced Engineering and Science (IAES) in collaboration with Intelektual Pustaka Media Utama (IPMU).
IJEECS visitor statistics