Hybrid Deepfake Detection: Leveraging ResNet50 and InceptionV3 for High-Accuracy Classification
Edidiong Akpabio, Arya Deshmukh, Supriya Narad · 2025
Deepfake technology has emerged as a significant issue because of its potential for misuse in the dissemination of false information, financial scams, and political manipulation. To counter this challenge, this paper introduces a deepfake detection method using a hybrid deep learning model that integrates ResNet50 and InceptionV3. The suggested model has a detection accuracy of 94%, proving its efficacy in separating real and fake media. The system is deployed in a Flutter-based smartphone app, making real-time deepfake detection possible. This paper outlines the dataset, training of the model, and performance measurements, emphasizing its real-world utility in security and media authenticity assurance.