Curvelet-Based Frequency-Aware Feature Enhancement for Deepfake Detection

TL;DR

This paper introduces a novel deepfake detection method using Curvelet transform to enhance frequency domain features, improving robustness against compression artifacts.

Contribution

It pioneers the use of Curvelet Transform in deepfake detection, integrating wedge-level attention and scale-aware masking for improved feature discrimination.

Findings

Achieves 98.48% accuracy on FF++ low compression dataset.

Maintains high performance under high compression conditions.

Demonstrates the effectiveness of Curvelet-based features in deepfake detection.

Abstract

The proliferation of sophisticated generative models has significantly advanced the realism of synthetic facial content, known as deepfakes, raising serious concerns about digital trust. Although modern deep learning-based detectors perform well, many rely on spatial-domain features that degrade under compression. This limitation has prompted a shift toward integrating frequency-domain representations with deep learning to improve robustness. Prior research has explored frequency transforms such as Discrete Cosine Transform (DCT), Fast Fourier Transform (FFT), and Wavelet Transform, among others. However, to the best of our knowledge, the Curvelet Transform, despite its superior directional and multiscale properties, remains entirely unexplored in the context of deepfake detection. In this work, we introduce a novel Curvelet-based detection approach that enhances feature quality through wedge-level attention and scale-aware spatial masking, both trained to selectively emphasize discriminative frequency components. The refined frequency cues are reconstructed and passed to a modified pretrained Xception network for classification. Evaluated on two compression qualities in the challenging FaceForensics++ dataset, our method achieves 98.48% accuracy and 99.96% AUC on FF++ low compression, while maintaining strong performance under high compression, demonstrating the efficacy and interpretability of Curvelet-informed forgery detection.