Back to Projects

Deep Learning

CNN

Attention Mechanisms

TensorFlow

Keras

OpenCV

Python

Deepfake Detection using CNN with Attention Mechanisms

Comparative study of CNN architectures enhanced with SE Attention and CBAM for binary classification of authentic vs. AI-generated facial images. Achieved 91.7% validation accuracy with the SE Attention model.

Base CNN

accuracy82.4%

precision80.1%

recall84.3%

F1 Score82.2%

CNN + SE Attention

accuracy91.7%

precision90.2%

recall93.1%

F1 Score91.6%

CNN + CBAM

accuracy88.3%

precision87%

recall89.5%

F1 Score88.2%

Model Comparison