Next-Frame Feature Prediction for Multimodal Deepfake Detection and Temporal Localization
This work addresses the challenge of detecting deepfakes in videos, which is crucial for security and media integrity, by improving generalization and localization capabilities, though it is incremental in nature.
The paper tackles the problem of multimodal deepfake detection and temporal localization by proposing a single-stage training framework that incorporates next-frame prediction and a window-level attention mechanism, achieving strong generalization and precise localization across multiple benchmark datasets.
Recent multimodal deepfake detection methods designed for generalization conjecture that single-stage supervised training struggles to generalize across unseen manipulations and datasets. However, such approaches that target generalization require pretraining over real samples. Additionally, these methods primarily focus on detecting audio-visual inconsistencies and may overlook intra-modal artifacts causing them to fail against manipulations that preserve audio-visual alignment. To address these limitations, we propose a single-stage training framework that enhances generalization by incorporating next-frame prediction for both uni-modal and cross-modal features. Additionally, we introduce a window-level attention mechanism to capture discrepancies between predicted and actual frames, enabling the model to detect local artifacts around every frame, which is crucial for accurately classifying fully manipulated videos and effectively localizing deepfake segments in partially spoofed samples. Our model, evaluated on multiple benchmark datasets, demonstrates strong generalization and precise temporal localization.