Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{186445,
        author = {Akshay Jadhav and Om Mangate and Unmesh Kakuste and Satyam Shinde},
        title = {CNN Ensemble Model for Real-Time Deepfake Image Detection},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {6},
        pages = {1331-1343},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=186445},
        abstract = {The proliferation of the ‘Deepfakes’ is increasingly undermining media authenticity and digital security. This work offers an ensemble-based deepfake detection system using EfficientNetB1, ResNet50, and Xception models. Every architecture is fine-tuned on a labelled dataset of actual and fake facial images; then, using a soft voting technique, they are combined to enhance classification robustness. Our method shows the benefit of architectural variety by attaining greater accuracy and AUC over single models. Moreover, GradCAM visualisations are used to interpret predictions by localising facial areas affecting model decisions. The suggested approach shows good generalisation ability and provides a scalable and understandable solution for deepfake image detection in the real world.},
        keywords = {Ensemble, EfficientNetB1, ResNet50, and Xception, Grad CAM, Media Forensics.},
        month = {November},
        }