SpectraNet: FFT-assisted Deep Learning Classifier for Deepfake Face Detection

TL;DR

The paper tackles robust, scalable detection of facial deepfakes under severe class imbalance. It proposes a lightweight binary classifier based on EfficientNet-B6, trained with oversampling and robust preprocessing, while also evaluating a Fourier-transform frequency model that shows limited practical benefit. The approach achieves high performance with accuracy and AUC around $0.91$, and demonstrates faster inference than the Fourier-hybrid variant on modern GPUs. Overall, it contributes a practical, generalizable solution for deepfake detection that balances accuracy, efficiency, and generalization across datasets.

Abstract

Detecting deepfake images is crucial in combating misinformation. We present a lightweight, generalizable binary classification model based on EfficientNet-B6, fine-tuned with transformation techniques to address severe class imbalances. By leveraging robust preprocessing, oversampling, and optimization strategies, our model achieves high accuracy, stability, and generalization. While incorporating Fourier transform-based phase and amplitude features showed minimal impact, our proposed framework helps non-experts to effectively identify deepfake images, making significant strides toward accessible and reliable deepfake detection.

Figures (1)

• Figure 1: Transformed instances of input images, including augmentations and preprocessing steps