aiqbench analysis
This research introduces CTCLD, a novel framework that leverages cross-modal translation between audio and visual signals to enhance video deepfake detection. By establishing a Bayesian foundation for decomposing joint distributions, the method effectively captures subtle inconsistencies in manipulated signals through bidirectional latent denoising. The approach demonstrates robust performance by aligning heterogeneous modalities in latent spaces, addressing critical gaps where previous methods failed to utilize intrinsic dependencies between audio and visual data.
Cross-modal Translation via Conditional Latent Denoising for Video Deepfake Detection
arXiv:2609.33394v1 Announce Type: new Abstract: The growing threat of video deepfakes necessitates multimodal detection. Beyond serving as independent indicators of authenticity, audio and visual signals have intrinsic dependencies that also provide an essential criterion for detection. Previous methods often overlook the cross-modal correspondences, hindering information transfer between domains and leaving crucial detection cues unexplored. To address this challenge, we propose a framework called Cross-modal Translation via Conditional Latent Denoising (CTCLD) for video deepfake detection. It connects the distinct distributions of heterogeneous modalities in latent spaces, enabling smooth cross-domain information transfer to improve detection performance. We first establish a Bayesian foundation by decomposing the audio-visual joint distribution. Subsequently, CTCLD translates both modalities via bidirectional latent denoising conditioned on each other, effectively capturing subtle inconsistencies in the manipulated signals. Experimental results demonstrate that the proposed CTCLD enables comprehensive domain alignment, resulting in a robust video deepfake detection approach with competitive performance.
Abstract:The growing threat of video deepfakes necessitates multimodal detection. Beyond serving as independent indicators of authenticity, audio and visual signals have intrinsic dependencies that also provide an essential criterion for detection. Previous methods often overlook the cross-modal correspondences, hindering information transfer between domains and leaving crucial detection cues unexplored. To address this challenge, we propose a framework called Cross-modal Translation via Conditional Latent Denoising (CTCLD) for video deepfake detection. It connects the distinct distributions of heterogeneous modalities in latent spaces, enabling smooth cross-domain information transfer to improve detection performance. We first establish a Bayesian foundation by decomposing the audio-visual joint distribution. Subsequently, CTCLD translates both modalities via bidirectional latent denoising conditioned on each other, effectively capturing subtle inconsistencies in the manipulated signals. Experimental results demonstrate that the proposed CTCLD enables comprehensive domain alignment, resulting in a robust video deepfake detection approach with competitive performance.
Artificial Intelligence (cs.AI)
arXiv:2609.33394 [cs.AI]
(or arXiv:2609.33394v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.33394 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 238 words
Related coverage
VSA: Accelerating Video Diffusion Inference with Sparse Attention on AMD GPUs
AMD · Aug 4, 2026
Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AWS · Sep 4, 2026
Synthesize Realistic 3D Medical Images at Scale to Ship Pre‑Trained Models
NVIDIA · May 22, 2026
Runway’s WorldPrompt and the Engineering of Real-Time Worlds
Latent Space · Sep 25, 2026