# ForensicFlow: A Tri-Modal Adaptive Network for Robust Deepfake Detection

arXiv: https://arxiv.org/abs/2511.14554 · HTML: https://mlsanity.com/paper/2511.14554/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 5 of its 9 references that are in the map.
- First posted on arXiv 18 Nov 2025.

## Builds on (most cited first, 5 of 5)

- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [Celeb-DF: A Large-scale Challenging Dataset for DeepFake Forensics](https://mlsanity.com/paper/1909.12962/) (2019)
- [PUDD: Towards Robust Multi-modal Prototype-based Deepfake Detection](https://mlsanity.com/paper/2406.15921/) (2024)
- [FSBI: Deepfakes Detection with Frequency Enhanced Self-Blended Images](https://mlsanity.com/paper/2406.08625/) (2024)

## Abstract (arXiv)

Modern deepfakes evade detection by leaving subtle, domain-speci c artifacts that single branch networks miss. ForensicFlow addresses this by fusing evidence across three forensic dimensions: global visual inconsistencies (via ConvNeXt-tiny), ne-grained texture anomalies (via Swin Transformer-tiny), and spectral noise patterns (via CNN with channel attention). Our attention-based temporal pooling dynamically prioritizes high-evidence frames, while adaptive fusion weights each branch according to forgery type. Trained on CelebDF(v2) with Focal Loss, the model achieves AUC 0.9752, F1 0.9408, and accuracy 0.9208 out performing single-stream detectors. Ablation studies con rm branch synergy, and Grad-CAM visualizations validate focus on genuine manipulation regions (e.g., facial boundaries). This multi-domain fusion strategy establishes robustness against increasingly sophisticated forgeries.