# Robust Deepfake Detection: Mitigating Spatial Attention Drift via Calibrated Complementary Ensembles

arXiv: https://arxiv.org/abs/2604.25889 · HTML: https://mlsanity.com/paper/2604.25889/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 39 references that are in the map.
- First posted on arXiv 28 Apr 2026.

## Builds on (most cited first, 25 of 25)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution](https://mlsanity.com/paper/2202.10054/) (2022)
- [Transformer Interpretability Beyond Attention Visualization](https://mlsanity.com/paper/2012.09838/) (2020)
- [Score-CAM: Score-Weighted Visual Explanations for Convolutional Neural Networks](https://mlsanity.com/paper/1910.01279/) (2019)
- [Celeb-DF: A Large-scale Challenging Dataset for DeepFake Forensics](https://mlsanity.com/paper/1909.12962/) (2019)
- [Leveraging Frequency Analysis for Deep Fake Image Recognition](https://mlsanity.com/paper/2003.08685/) (2020)
- [Towards Universal Fake Image Detectors that Generalize Across Generative Models](https://mlsanity.com/paper/2302.10174/) (2023)
- [Towards Real-World Blind Face Restoration with Generative Facial Prior](https://mlsanity.com/paper/2101.04061/) (2021)
- [What Makes Training Multi-Modal Classification Networks Hard?](https://mlsanity.com/paper/1905.12681/) (2019)
- [Detecting Deepfakes with Self-Blended Images](https://mlsanity.com/paper/2204.08376/) (2022)
- [DeeperForensics-1.0: A Large-Scale Dataset for Real-World Face Forgery Detection](https://mlsanity.com/paper/2001.03024/) (2020)
- [The Deepfake Detection Challenge (DFDC) Preview Dataset](https://mlsanity.com/paper/1910.08854/) (2019)
- [The Shattered Gradients Problem: If resnets are the answer, then what is the question?](https://mlsanity.com/paper/1702.08591/) (2017)
- [FaceForensics: A Large-scale Video Dataset for Forgery Detection in Human Faces](https://mlsanity.com/paper/1803.09179/) (2018)
- [LAA-Net: Localized Artifact Attention Network for Quality-Agnostic and Generalizable Deepfake Detection](https://mlsanity.com/paper/2401.13856/) (2024)
- [Face Forensics in the Wild](https://mlsanity.com/paper/2103.16076/) (2021)
- [Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector](https://mlsanity.com/paper/2503.20188/) (2025)
- [Towards More General Video-based Deepfake Detection through Facial Component Guided Adaptation for Foundation Model](https://mlsanity.com/paper/2404.05583/) (2024)
- [Exploring Unbiased Deepfake Detection via Token-Level Shuffling and Mixing](https://mlsanity.com/paper/2501.04376/) (2025)
- [Celeb-DF++: A Large-scale Challenging Video DeepFake Benchmark for Generalizable Forensics](https://mlsanity.com/paper/2507.18015/) (2025)
- [Robust Deepfake Detection, NTIRE 2026 Challenge: Report](https://mlsanity.com/paper/2604.24163/) (2026)
- [Practical Manipulation Model for Robust Deepfake Detection](https://mlsanity.com/paper/2506.05119/) (2025)
- [DDL: A Large-Scale Dataset for Deepfake Detection and Localization in Diversified Real-World Scenarios](https://mlsanity.com/paper/2506.23292/) (2025)
- [Patch-Discontinuity Mining for Generalized Deepfake Detection](https://mlsanity.com/paper/2512.22027/) (2025)

## Abstract (arXiv)

Current deepfake detection models achieve state-of-the-art performance on pristine academic datasets but suffer severe spatial attention drift under real-world compound degradations, such as blurring and severe lossy compression. To address this vulnerability, we propose a foundation-driven forensic framework that integrates an extreme compound degradation engine with a structurally constrained, multi-stream architecture. During training, our degradation pipeline systematically destroys high-frequency artifacts, optimizing the DINOv2-Giant backbone to extract invariant geometric and semantic priors. We then process images through three specialized pathways: a Global Texture stream, a Localized Facial stream, and a Hybrid Semantic Fusion stream incorporating CLIP. Through analyzing spatial attribution via Score-CAM and feature stability using Cosine Similarity, we quantitatively demonstrate that these streams extract non-redundant, complementary feature representations and stabilize attention entropy. By aggregating these predictions via a calibrated, discretized voting mechanism, our ensemble successfully suppresses background attention drift while acting as a robust geometric anchor. Our approach yields highly stable zero-shot generalization, achieving Fourth Place in the NTIRE 2026 Robust Deepfake Detection Challenge at CVPR. Code is available at https://github.com/khoalephanminh/ntire26-deepfake-challenge.