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{185950,
        author = {Saloni Gawande and Dr.Nilesh S. Wadhe and Pruthvi S Dharmale and Gauri A Dharao and Dipak V Botke},
        title = {Cognitive Image Classifier},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {5},
        pages = {3338-3344},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=185950},
        abstract = {Artificial intelligence and generative models are making it easier to create modified and false images one can find online. These technologies are helpful for being creative and come up with new ideas, but they come with the risk of trolling and spreading deceiving data, which might lower people's trust in digital content. For photo verification, generative models like GANs and deepfake algorithms are often more advanced than traditional methods. The capability of intelligent systems to tell the difference between real, altered, and computer-generated images thus grows more and more essential. To tackle this issue, the Cognitive Image Classifier project proposal develops a framework utilizing deep learning for identifying images into the following groups: synthetic-fake, real-manipulated.
The system leverages cognitive computing concepts for decision making, similar to that of human judgment, by incorporating artificial neural networks. Prior to being introduced into the classification framework, the images are pre-processed using techniques such as noise filtering, histogram equalization, and edge detection, which helps to ensure detection of even the slightest alterations or obfuscated manipulations. The framework utilizes Convolutional Neural Networks (CNNs), and benefits from transfer learning from multiple pre-trained models including Inception and VGG-16. The proposed system development will evaluate opportunities for hardware projects that consider adaptations for deployment but also recognize the fidelity and capabilities of offered computational resources.  Ultimately, this research and project will contribute toward improvements in useable technology and recognize shifts in trust and credibility in digitally formatted images.},
        keywords = {Cognitive Computing, Image Classification, Deep Learning, Convolutional Neural Network (CNN), Transfer Learning, Natural Images, Synthetic Images, Real vs Fake Detection, Image Forensics, Deepfake Detection.},
        month = {October},
        }