• Version
• Download 8
• File Size 630.17 KB
• File Count 1
• Create Date 22 March 2026
• Last Updated 22 March 2026

Deepfake Video Detection using EfficientNet and CNN-LSTM Architecture

RajeshKannan T
Department of CSE,Dr. M.G.R Educational and Research Institute, Maduravoyal, Chennai
rajeshkannan1572005@gmail.com
Mrs. Revathi
Assistant Professor,Department of CSE,Dr. M. G. R Educational and ResearchInstitute,Maduravoyal, Chennai
Ramarathinam M
Department of CSE,Dr. M.G.R Educational and ResearchInstitute,Maduravoyal, Chennai ramrathinam2327@gmail.com
Dr. T. Kumanan
Professor,Department of CSE,Dr. M. G. R Educational and ResearchInstitute,Maduravoyal, Chennai
kumanan.cse@drmgr.ac.in

Moulidharan AD
Department of CSE,Dr. M.G.R Educational and Research Institute,Maduravoyal, Chennai
moulidharan.a.d@gmail.com
Dr. M. Nisha
Professor,Department of CSE,Dr. M. G. R Educational and ResearchInstitute,Maduravoyal, Chennai
nisha.cse@drmgrdu.ac.in

Abstract—Deepfake videos generated using advanced artificial intelligence techniques have become increasingly realistic and pose serious threats to digital media authenticity and public trust. Detecting manipulated videos is therefore an important challenge in modern computer vision research. Many existing deepfake detection approaches analyze complete video frames using complex deep learning models, which increasescomputational cost and processing time. To address this issue,  this work proposes a deepfake video detection approachthatfocuses mainly on facial regions, sincemostdeepfakemanipulationsoccurontheface.DeepfakeDetection,DeepLearning,onvolutionalNeural Network (CNN), EfficientNet, Face Detection, Frame Extraction, Facial Feature Analysis, Video Manipulation Detection, Image Processing, Artificial Intelligence