Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
CAFNet, a novel 576k-parameter architecture, jointly tackles ternary classification (real, fully-fake, or half-truth audio) and temporal boundary regression of manipulated regions in audio deepfakes. It fuses MFCC, LFCC, and Chroma-STFT features using cross-attentive depthwise-separable convolutions and a BiLSTM regression head. CAFNet achieves state-of-the-art performance on the MLADDC dataset, outperforming much larger models (XLS-R 300M, AST 87M) in binary detection with significantly fewer parameters and demonstrating superior cross-domain generalization.
You can now detect and localize spliced audio deepfakes with a tiny model that crushes larger, fine-tuned alternatives and generalizes better across datasets.
Audio deepfake detection is well-studied as a binary problem, but partially manipulated speech, where a short synthesised segment is spliced into an otherwise genuine utterance, poses a harder and more realistic threat. Detecting such half-truth audio requires not only distinguishing it from real and fully fake speech, but also localising where the manipulation occurs. We present CAFNet, a 576k-parameter architecture that addresses both tasks jointly: it performs ternary classification (real, fully-fake, or half-truth) and regresses the temporal boundaries of the synthesised region in a single forward pass. CAFNet fuses Mel-Frequency Cepstral Coefficient (MFCC), Linear-Frequency Cepstral Coefficient (LFCC), and Chroma Short-Time Fourier Transform (Chroma-STFT) features through parallel depthwise-separable convolution branches with cross-attention, followed by a Bidirectional Long Short-Term Memory (BiLSTM) regression head for boundary prediction. On the combined Multi-Lingual Audio Deepfake Detection Corpus (MLADDC) T2+T3 test set, CAFNet achieves 92.71% accuracy and macro Area Under the Curve (AUC) of 0.9910, with boundary localisation Mean Absolute Error (MAE) of 0.075s and a median error of 0.052s. On binary detection, it achieves 96.76% accuracy and 3.20% Equal Error Rate (EER), outperforming fine-tuned XLS-R 300M (78.31%) and AST 87M (93.03%) at over 500 times fewer parameters. A cross-dataset study further shows that standard fine-tuning collapses cross-domain representations even under reduced backbone learning rates.