Loading...
|
Please use this identifier to cite or link to this item:
http://asiair.asia.edu.tw/ir/handle/310904400/115524
|
| Title: | AI-enabled digital forgery analysis and crucial interactions monitoring in smart communities |
| Authors: | Sedik, Ahmed;Sedik, Ahmed;Male, Yassine;Maleh, Yassine;El, Ghada M.;Banby, Ghada M. El;Ashraf, Ashraf A.M.;Khalaf, Ashraf A.M.;Abd, Fathi E.;Fathi, E. Abd El-Samie;Bhoosha, Brij;Gupta, Brij Bhooshan;Konstantinos;Psannis, Konstantinos;Abd, Ahmed A.;Ahmed, Ahmed A. Abd El-Latif |
| Contributors: | 資訊電機學院資訊工程學系 |
| Keywords: | Forgery detection；Deep learning；IoT；Smart cities；Security analysis |
| Date: | 2022-04-01 |
| Issue Date: | 2023-03-29 02:30:30 (UTC+0) |
| Publisher: | 亞洲大學 |
| Abstract: | Digital forgery has become one of the attractive research fields in today’s technology. There are several types of forgery in digital media transmission, especially digital image transmission. A common type of forgery is copy-move forgery (CMF). The CMF may be encountered in streets, railway stations, underground stations, or festivals. This type of forgery may lead to hugger-mugger in some cases. Therefore, there is a need to find a sufficient countermeasure mechanism to detect image forgeries. This paper presents a new CMFD approach that depends on deep learning for IoT based smart cities. Two well-known deep learning models, namely CNN and ConvLSTM, are adopted for CMFD. The proposed models are tested on MICC-220, MICC-600 and MICC 2000 datasets for validation. Several tests are performed to verify the effectiveness of the proposed models. The simulation results reveal that the testing accuracy reaches 95%, 73%, and 94% for MICC-F220, MICC-F600 and MICC-F2000 datasets. In addition, the proposed approach achieves an accuracy of 85% for a combined set of all datasets. |
| Appears in Collections: | [資訊工程學系] 期刊論文
|
Files in This Item:
| File |
Description |
Size | Format | |
|---|
| index.html | | 0Kb | HTML | 97 | View/Open |
|
All items in ASIAIR are protected by copyright, with all rights reserved.
|
