A Comprehensive Survey on Image Forgery Detection Techniques Using Handcrafted and Deep Features

Despite significant advancements in image forgery detection, several open challenges and research oppor-tunities remain, providing ample scope for future work. One of the primary directions is the development of generalizable detection frameworks capable of maintaining robust performance across diverse datasets, manipulation types, and post-processing operations. Many existing methods are trained and evaluated on specific datasets, leading to performance degradation when applied to unseen data distribu-tions. Future research should emphasize domain adaptation, cross-dataset evaluation, and self-supervised learning to improve generalization. Another important research avenue lies in robust detection under advanced and emerging forgery techniques. With the rapid evolution of generative models, particularly Generative Adversarial Networks (GANs) and diffusion-based image synthesis, forgery artifacts are becoming increasingly subtle and difficult to detect. Future detection systems must incorporate multi-scale analysis, frequency-domain cues, and semantic consistency modeling to effectively counter sophisticated deepfake and AI-generated content. The integration of hybrid and multimodal feature learning also presents a promising direction. Combining spatial, frequency, noise, and semantic features with deep representations can further enhance robustness against complex manipulation pipelines. Additionally, leveraging multimodal data such as metadata, camera sensor information, and temporal cues from image sequences may significantly improve detection reliability in real-world forensic applications. Lightweight and computationally efficient models are essential for deploying image forgery de-tection systems on resource-constrained platforms, such as mobile devices and edge computing environ-ments. Future work should explore model compression, knowledge distillation, and efficient transformer architectures to balance detection accuracy with computational efficiency. Another critical aspect is the need for standardized datasets, benchmarks, and evaluation protocols. Although several datasets exist, they vary widely in manipulation types, quality, and anno-tation standards. Future efforts should focus on creating large-scale, diverse, and continuously updated benchmark datasets that reflect real-world forgery scenarios. Moreover, establishing unified evaluation metrics and testing protocols will enable fair and reproducible comparisons across methods. Finally, improving the interpretability and explainability of image forgery detection models is crucial, especially for applications in legal, forensic, and journalistic domains. Future research should emphasize explainable AI (XAI) techniques that provide visual and quantitative evidence supporting model decisions, thereby increasing user trust and practical usability. In summary, future research in image forgery detection should aim to develop robust, explainable, efficient, and generalizable frameworks capable of addressing the growing challenges posed by advanced image manipulation techniques. These advancements will play a vital role in ensuring the integrity and trustworthiness of visual media in an increasingly digital world.  

[1]         J. Fridrich, D. Soukal, and J. Lukas, “Detection of Copy-Move Forgery in Digital Images,” in Proc. Digital Forensic Research Workshop, 2003. [2]         A. C. Popescu and H. Farid, “Exposing Digital Forgeries by Detecting Duplicated Image Regions,” IEEE Trans. Signal Processing, vol. 53, no. 2, pp. 758–767, 2004. [3]         J. Lukas, J. Fridrich, and M. Goljan, “Detecting Digital Image Forgeries Using Sensor Pattern Noise,” in Proc. SPIE, 2006. [4]         T. T. Ng, S. Chang, and Q. Sun, “Blind Detection of Photomontage Using Higher Order Statistics,” in Proc. IEEE ISCAS, 2009. [5]         I. Amerini, L. Ballan, R. Caldelli, and A. Del Bimbo, “A SIFT-Based Forensic Method for Copy-Move Attack Detection,” IEEE Trans. Inf. Forensics Security, vol. 6, no. 3, pp. 1099–1110, 2011. [6]         J. Dong, W. Wang, and T. Tan, “CASIA Image Tampering Detection Evaluation Database,” in Proc. IEEE ChinaSIP, 2011. [7]         V. Christlein, C. Riess, J. Jordan, C. Riess, and E. Angelopoulou, “An Evaluation of Popular Copy-Move Forgery Detection Approaches,” IEEE Trans. Inf. Forensics Security, vol. 7, no. 6, pp. 1841–1854, 2012. [8]         B. Bayar and M. C. Stamm, “A Deep Learning Approach to Universal Image Manipulation Detec-tion,” IEEE Trans. Inf. Forensics Security, vol. 12, no. 11, pp. 2491–2503, 2017. [9]         Y. Rao and J. Ni, “A Deep Learning Approach to Detection of Splicing and Copy-Move Forgeries,” in Proc. IEEE WIFS, 2017. [10]      L. Verdoliva, “Media Forensics: An Overview,” IEEE Signal Processing Magazine, vol. 35, no. 1, pp. 16–28, 2018.