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{198612,
        author = {KURAKULA VARALAKSHMI and LELLA KISHAN CHANDRA DEV and MADALA PRAVEENA and MARUBOYINA VAMSI and DR.D.HEMA},
        title = {REAL/FAKE LOGO AND DEEP FAKE IMAGE DETECTION SYSTEM},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {8484-8488},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=198612},
        abstract = {The rapid proliferation of sophisticated generative modeling techniques has introduced unprecedented challenges to digital authenticity, necessitating the development of robust, automated verification frameworks. This research presents a comprehensive Real/Fake Logo and Deepfake Image Detection System designed to safeguard brand integrity and combat the spread of hyper-realistic synthetic media. As generative adversarial networks (GANs) and diffusion models evolve, the distinction between authentic visual data and manipulated content becomes increasingly indistinguishable to the human eye, leading to significant risks in corporate security, misinformation, and digital forensics.
Our proposed system employs a dual-stream architectural approach to address these distinct yet overlapping threats. The first module focuses on Logo Authentication, utilizing a combination of Feature Pyramid Networks (FPN) and Siamese Neural Networks to identify subtle inconsistencies in geometry, color distribution, and spatial positioning that characterize counterfeit brand marks. By leveraging a high-resolution dataset of authentic corporate identities against diverse adversarial samples, the system achieves high sensitivity in detecting "brand-jacking" attempts.
The second module targets Deepfake Image Detection, utilizing a multi-scale Convolutional Neural Network (CNN) integrated with an attention mechanism to capture mesoscopic properties and frequency-domain anomalies often left behind by synthesis algorithms. Unlike traditional methods that rely on specific artifacts, our model analyzes the biological inconsistencies in facial features and the statistical distribution of pixel gradients to ensure generalized performance across various generation techniques.},
        keywords = {Deep Learning, Convolutional Neural Networks, Deepfake Detection, Logo Verification, Vision Transformers, Digital Forensics.},
        month = {April},
        }