CVPR_QTFP
Beyond [CLS] Token: Query-Driven Token-Level Forgery Purification for Generalizable Deepfake Detection
[CVPR 2026]
Authors: Changshuo Wang1,2,*, Jiangming Wang2,*, Ke-Yue Zhang2, Taiping Yao2, Shunli Wang2, Shouhong Ding2, Ran Yi1, Lizhuang Ma1
Institution: 1Shanghai Jiao Tong University, 2Tencent Youtu Lab
📢 News
- [2026-03] Our paper has been accepted by CVPR 2026!
- [2026-03] We have released the project repository. Code and models are coming soon.
💡 Abstract
We investigate state-of-the-art deepfake detectors that leverage ViT-based vision foundation models and discover that the [CLS] token suffers from the Pre-trained Information Bias (PIB), i.e., it tends to mainly focus on global semantics due to the knowledge dominated by pre-trained model parameters, while struggling to emphasize subtle local forgery cues. To overcome this limitation, one potential way is incorporating the token-level features to reform a new detection-specific token. To this end, we propose Query-Driven Token-Level Forgery Purification (QTFP) framework to better capture local forgery traces without losing useful pre-trained prior. Specifically, we first introduce randomly initialized, learnable query tokens independent of the backbone and prior knowledge, which can effectively aggregate multi-patch evidence into a global token for detection. To make query tokens focus on meaningful regions, we propose a theoretical fake-likelihood contrastive learning loss, which employs a weighting strategy to highlight significant fake regions while diminishing the impact of real-like patches. Using SNR theory, we verify that the designed weight is both reliable and informative. To further maintain useful authentic information, a real-attention alignment constraint is applied to query tokens. These designs go beyond relying solely on the [CLS] token by jointly reorganizing real and fake information across all tokens, which successfully enhance detector robustness. Extensive experiments on diverse datasets demonstrate the effectiveness of our method.
🛠️ Methodology
Figure 1: Overview of the QTFP framework. We use learnable query tokens to identify and purify forged regions at the token level.
📜 Citation
If you find our work useful, please cite:
@inproceedings{wang2026qtfp,
title={Beyond [CLS] Token: Query-Driven Token-Level Forgery Purification for Generalizable Deepfake Detection},
author={Wang, Changshuo and Wang, Jiangming and Zhang, Ke-Yue and Yao, Taiping and Wang, Shunli and Ding, Shouhong and Yi, Ran and Ma, Lizhuang},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2026}
}