DETECTION AND LOCALISATION OF DEEPFAKES USING DEEP LEARNING

COMPUTER SCIENCE ENGINEERING

In today’s digital era, the proliferation of deepfake technology has raised significant concerns about the authenticity and integrity of multimedia content. Deepfake involves the creation of highly realistic fake images and videos and misuse for spreading fake news, defaming individuals, and posing a significant threat to the integrity of digital content. To combat this challenge, our project “Deep learning approaches for robust deepfake detection and localization “aims to address this critical issue by developing a robust system for the identification and localization of deepfake content by using the DenseNet121 model. Densenet’s dense connections contribute to improved gradient flow, enabling better information propagation through the network leading to better classification results. The accurate identification of tampered regions holds the utmost importance for identifying the intentions of the offenders. For localization purposes, we use binary masking by segmenting the image with a specific threshold value and indicating the morphed regions. This proposed framework seamlessly integrates forgery detection and localization.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.