A Trustworthy Explainable Deep Learning Framework for Copy-Move Forgery Detection Using Multi-Level XAI Techniques

Copy-Move Forgery occurs when copies from different portions of images are placed at other locations to conceal or add content. This method is one of the most common and challenging types of digital manipulations. CNNs have shown great performance in detecting copy-move forgery; however, the issue with them lies in the fact that they are a “black-box.” Therefore, there is always suspicion about the results obtained by them in sensitive applications. We proposed an Explainable Artificial Intelligence (XAI) framework that uses our proposed detector alongside three well-known interpretability techniques: Grad-CAM, LIME, and SHAP. Our model not only correctly detects the forged region within the image but also offers understandable reasons behind the result obtained for the forgery location. Experiments on CASIA and CoMoFoD datasets show that we achieved 97.8%, 98.6%, 97.1%, and 97.8% accuracy, precision, recall, and F1-score, respectively, along with 0.72 IoU using Grad-CAM.

Copy-Move Forgery Detection · CNN · Explainable AI · Grad-CAM · LIME · SHAP · Digital Image Forensics

1. [1]. Fridrich, J., Soukal, D., & Lukáš, J. (2003). Detection of Copy-Move Forgery in Digital Images. DFRWS, pp. 55–61. [Google Scholar] [Crossref]

2. [2]. Popescu, A. C., & Farid, H. (2004). Exposing Digital Forgeries by Detecting Duplicated Image Regions. TR2004-515, Dartmouth College. [Google Scholar] [Crossref]

3. [3]. Amerini, I. et al. (2011). A SIFT-based forensic method for copy–move attack detection. IEEE Trans. Inf. Forensics Security, 6(3), 1099–1110. [Google Scholar] [Crossref]

4. [4]. Shivakumar, B. L., & Santhosh Baboo, S. (2011). Detection of region duplication forgery using SURF. IJCSI, 8(4), 199–205. [Google Scholar] [Crossref]

5. [5]. Ryu, S.-J. et al. (2010). Detection of copy-rotate-move forgery using Zernike moments. IH Conf., LNCS 6387, pp. 51–65. [Google Scholar] [Crossref]

6. [6]. Li, J. et al. (2015). Segmentation-based image copy-move forgery detection. IEEE Trans. Inf. Forensics Security, 10(3), 507–518. [Google Scholar] [Crossref]

7. [7]. Rao, Y., & Ni, J. (2016). A Deep Learning Approach to Detection of Splicing and Copy-Move Forgeries. IEEE WIFS, pp. 1–6. [Google Scholar] [Crossref]

8. [8]. Liu, Y. et al. (2019). Adversarial learning for constrained image splicing detection. IEEE Trans. Inf. Forensics Security, 14(10), 2551–2566. [Google Scholar] [Crossref]

9. [9]. Wu, Y. et al. (2018). BusterNet: Detecting Copy-Move Image Forgery. ECCV, LNCS 11210, pp. 170–186. [Google Scholar] [Crossref]

10. [10]. Zhong, J.-L. et al. (2020). Robust Copy-Move Forgery Detection by False Alarm Control. IEEE Trans. Multimedia, 22(9), 2285–2298. [Google Scholar] [Crossref]

11. [11]. Zhang, Y. et al. (2016). Image Region Forgery Detection: A Deep Learning Approach. SG-CRC, vol. 14, pp. 1–11. [Google Scholar] [Crossref]

12. [12]. Wang, Z. et al. (2023). Self-Supervised Vision Transformer for Image Forgery Detection. CVPR, pp. 14396–14405. [Google Scholar] [Crossref]

13. [13]. Liu, X. et al. (2022). PSCC-Net: Progressive Spatio-Channel Correlation Network. IEEE Trans. CSVT, 32(11), 7505–7517. [Google Scholar] [Crossref]

14. [14]. Zhao, H. et al. (2021). Multi-Attentional Deepfake Detection. CVPR, pp. 2185–2194. [Google Scholar] [Crossref]

15. [15]. Selvaraju, R. R. et al. (2017). Grad-CAM: Visual Explanations from Deep Networks. ICCV, pp. 618–626. [Google Scholar] [Crossref]

16. [16]. Chattopadhay, A. et al. (2018). Grad-CAM++: Generalized Gradient-Based Visual Explanations. WACV, pp. 839–847. [Google Scholar] [Crossref]

17. [17]. Ribeiro, M. T. et al. (2016). 'Why Should I Trust You?': Explaining Any Classifier. KDD, pp. 1135–1144. [Google Scholar] [Crossref]

18. [18]. Lundberg, S. M., & Lee, S.-I. (2017). A Unified Approach to Interpreting Model Predictions. NeurIPS, vol. 30, pp. 4765–4774. [Google Scholar] [Crossref]

19. [19]. Holzinger, A. et al. (2019). Causability and explainability of AI in medicine. WIREs DMKD, 9(4), e1312. [Google Scholar] [Crossref]

20. [20]. Omeiza, D. et al. (2021). Smooth Grad-CAM++. ITSC, pp. 49–55. [Google Scholar] [Crossref]

21. [21]. Danilevsky, M. et al. (2020). A Survey of Explainable AI for NLP. AACL-IJCNLP, pp. 447–459. [Google Scholar] [Crossref]

22. [22]. Verma, V. et al. (2022). Explainability in Image Forgery Detection Using Grad-CAM. PReMI, LNCS 13102, pp. 301–311. [Google Scholar] [Crossref]

23. [23]. Bidgoli, A. et al. (2023). Towards Interpretable Deepfake Detection: LIME-Based Approach. JVCIR, 90, 103726. [Google Scholar] [Crossref]

24. [24]. Dong, J. et al. (2013). CASIA Image Tampering Detection Evaluation Database. ChinaSIP, pp. 422–426. [Google Scholar] [Crossref]

25. [25]. Tralic, D. et al. (2013). CoMoFoD – New Database for Copy-Move Forgery Detection. ELMAR, pp. 49–54. [Google Scholar] [Crossref]

• What the Desert Fathers Teach Data Scientists: Ancient Ascetic Principles for Ethical Machine-Learning Practice
• Comparative Analysis of Some Machine Learning Algorithms for the Classification of Ransomware
• Comparative Performance Analysis of Some Priority Queue Variants in Dijkstra’s Algorithm
• Transfer Learning in Detecting E-Assessment Malpractice from a Proctored Video Recordings.
• Dual-Modal Detection of Parkinson’s Disease: A Clinical Framework and Deep Learning Approach Using NeuroParkNet