用注意力引导的对比学习提升伪造人脸检测鲁棒性
DFCon: Attention-Driven Supervised Contrastive Learning for Robust Deepfake Detection
- 采用注意力驱动的监督对比损失，增强特征区分能力
- 多模型融合在验证集达95.83%准确率，有效应对真实场景
- 适合需要高鲁棒性深度伪造检测的应用场景
本文介绍了参加IEEE SP Cup 2025：野外深伪人脸检测（DFWild-Cup）的方案，聚焦于跨多样数据集的深伪人脸检测。方法采用MaxViT、CoAtNet和EVA-02等先进骨干网络，通过监督对比损失微调以增强特征分离。这些模型因互补优势被选中：MaxViT结合卷积与步进注意力，擅长捕捉局部特征；CoAtNet混合卷积与注意力机制，有效提取多尺度特征；EVA-02通过掩码图像建模进行强预训练，擅长捕获全局特征。训练后冻结模型参数，仅训练分类头，最终使用多数投票集成策略融合各模型预测，提升对未知场景的鲁棒性和泛化能力。系统在验证集上达到95.83%的准确率，有效应对真实世界中的深伪检测挑战。
原文摘要 · Abstract (English)
This report presents our approach for the IEEE SP Cup 2025: Deepfake Face Detection in the Wild (DFWild-Cup), focusing on detecting deepfakes across diverse datasets. Our methodology employs advanced backbone models, including MaxViT, CoAtNet, and EVA-02, fine-tuned using supervised contrastive loss to enhance feature separation. These models were specifically chosen for their complementary strengths. Integration of convolution layers and strided attention in MaxViT is well-suited for detecting local features. In contrast, hybrid use of convolution and attention mechanisms in CoAtNet effectively captures multi-scale features. Robust pretraining with masked image modeling of EVA-02 excels at capturing global features. After training, we freeze the parameters of these models and train the classification heads. Finally, a majority voting ensemble is employed to combine the predictions from these models, improving robustness and generalization to unseen scenarios. The proposed system addresses the challenges of detecting deepfakes in real-world conditions and achieves a commendable accuracy of 95.83% on the validation dataset.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。