Research Article
AI Vs Real Image Classification Using Deep Learning
@INPROCEEDINGS{10.4108/eai.28-4-2025.2358080, author={Lijetha C Jaffrin and Kolli. Anvesh and Yelluri. Balaji Venkata Ratnam and Vaddi. Nikhileshwar Reddy}, title={AI Vs Real Image Classification Using Deep Learning}, proceedings={Proceedings of the 4th International Conference on Information Technology, Civil Innovation, Science, and Management, ICITSM 2025, 28-29 April 2025, Tiruchengode, Tamil Nadu, India, Part II}, publisher={EAI}, proceedings_a={ICITSM PART II}, year={2025}, month={10}, keywords={deep learning ai-generated images image classification mobilenetv2 flask deployment digital image authentications real vs ai detection}, doi={10.4108/eai.28-4-2025.2358080} }- Lijetha C Jaffrin
Kolli. Anvesh
Yelluri. Balaji Venkata Ratnam
Vaddi. Nikhileshwar Reddy
Year: 2025
AI Vs Real Image Classification Using Deep Learning
ICITSM PART II
EAI
DOI: 10.4108/eai.28-4-2025.2358080
Abstract
With the rapid development of artificial intelligence (AI), synthetic images are being generated to an extent that it is now close to impossible to tell if an image is real or not. This paper describes a deep learning technique for classifying real vs. AI images based on the MobileNetV2 architecture. A dataset with 24,000 real and 24,000 AI-generated training images and 6,000 real and 6,000 AI-generated test images is employed for the model training and testing. Model is fine-tuned along with data augmentations and optimized by adaptive learning rate schedule. Experiments show the capability of the model for recognizing AI-generated images with high accuracy and efficient classification performance. The proposed system is further deployed using Flask library to make a web-based client where user can upload images and receive real-time predictions. The relevance of deep learning in solving digital authenticity problems and produce certifiable visual content.
