International Journal of Innovative Research in Computer and Communication Engineering
ISSN Approved Journal | Impact factor: 8.771 | ESTD: 2013 | Follows UGC CARE Journal Norms and Guidelines
| Monthly, Peer-Reviewed, Refereed, Scholarly, Multidisciplinary and Open Access Journal | High Impact Factor 8.771 (Calculated by Google Scholar and Semantic Scholar | AI-Powered Research Tool | Indexing in all Major Database & Metadata, Citation Generator | Digital Object Identifier (DOI) |
| TITLE | An Explainable AI Method for Detecting Fake Facial Images |
|---|---|
| ABSTRACT | The rapid advancement of generative adversarial networks and diffusion-based image synthesis has produced synthetic facial images visually indistinguishable from authentic photographs, creating serious threats to digital identity, media integrity, and public trust. Existing detection approaches suffer from two critical limitations: inadequate sensitivity to color-space artifacts when processing standard RGB images, and a black-box decision-making process that prevents forensic analysts and legal professionals from validating classification outcomes. This study addresses both limitations by proposing a deepfake facial image detection framework integrating MTCNN-based face detection and geometric alignment to ensure the classifier focuses exclusively on forensically relevant facial content rather than background noise, LAB color space transformation to decouple luminance from chrominance and surface subtle generation artifacts characteristic of synthesized faces, fine-tuned InceptionResNetV2 classification, and LIME-based superpixel attribution maps highlighting suspicious facial zones including skin-hair boundaries, periocular regions, and dentition areas. The framework was trained on the 140k Real and Fake Faces dataset using an 80/10/10 split on a GPU-enabled environment. The proposed pipeline achieved a test accuracy of 98.93%, generalization accuracy of 96.94%, with good precision, recall and F1-score, outperforming the baseline RGB model across every evaluated metric, with demonstrated real-time deployability through a Flask web application. |
| AUTHOR | VIGHNESH V PRABHU, NIKITHA K S Department of CS& E, Bangalore Institute of Technology, Bengaluru, India Assistant Professor, Department of CS& E, Bangalore Institute of Technology, Bengaluru, India |
| VOLUME | 186 |
| DOI | DOI: 10.15680/IJIRCCE.2026.1407064 |
| pdf/64_An Explainable AI Method for Detecting Fake Facial Images.pdf | |
| KEYWORDS | |
| References | [1] I. J. Goodfellow et al., “Generative Adversarial Nets,” in Proc. NeurIPS, vol. 27, pp. 2672–2680, 2014. [2] T. Karras et al., “Analyzing and Improving the Image Quality of StyleGAN,” in Proc. IEEE/CVF CVPR, pp. 8110–8119, 2020. [3] A. Rossler et al., “FaceForensics++: Learning to Detect Manipulated Facial Images,” in Proc. IEEE/CVF ICCV, pp. 1–11, 2019. [4] B. Dolhansky et al., “The Deepfake Detection Challenge (DFDC) Dataset,” arXiv:2006.07397, 2020. [5] M. A. Khan et al., “A Survey on Multimedia-Enabled Deepfake Detection: State-of-the-Art Tools and Techniques, Emerging Trends, Current Challenges and Future Directions,” IEEE Access, vol. 13, pp. 1–28, 2025. [6] A. Mathews et al., “An Explainable Deepfake Detection Framework on a Novel Unconstrained Dataset,” Expert Systems with Applications, vol. 213, p. 119030, 2023. [7] S. Silva et al., “Deepfake Forensics Analysis: An Explainable Hierarchical Ensemble of Weakly Supervised Models,” Forensic Science International: Digital Investigation, vol. 42, p. 301390, 2022. [8] S. Cirillo et al., “Explainability-Driven Adversarial Robustness Assessment for Generalized Deepfake Detectors,” Pattern Recognition Letters, vol. 178, pp. 45–52, 2025. [9] U. Mansoor and A. Iliev, “Explainable AI for DeepFake Detection,” Journal of Imaging, vol. 11, no. 2, pp. 1–18, 2025. [10] K. Barik et al., “Practical Evaluation and Performance Analysis for Deepfake Detection Using Advanced AI Models,” Multimedia Tools and Applications, vol. 83, pp. 12541–12563, 2024. [11] A. Soudy et al., “Deepfake Detection Using Convolutional Vision Transformers and CNNs,” Neural Computing and Applications, vol. 36, pp. 8741–8756, 2024. [12] M. Abu Talib et al., “Chrominance and Luminance: A Study to Detect Deepfakes,” Journal of King Saud University – CIS, vol. 37, no. 1, pp. 1–14, 2025. [13] M. S. Momin et al., “Explainable Deepfake Detection Across Different Modalities: An Overview of Methods and Challenges,” Artificial Intelligence Review, vol. 58, no. 3, pp. 1–42, 2025. [14] K. Zhang et al., “Joint Face Detection and Alignment Using Multi-Task Cascaded Convolutional Networks,” IEEE Signal Processing Letters, vol. 23, no. 10, pp. 1499–1503, 2016. [15] M. T. Ribeiro et al., “‘Why Should I Trust You?’: Explaining the Predictions of Any Classifier,” in Proc. 22nd ACM SIGKDD, pp. 1135–1144, 2016. [16] C. Szegedy et al., “Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning,” in Proc. AAAI, pp. 4278–4284, 2017. |
