$μ$Flow: Leveraging Average Images for Improving Generalisation of Deepfake Faces Detectors
2026-06-29 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionMachine Learning
AI summaryⓘ
The authors created a new deepfake detector called μFlow that only learns from real images, unlike other detectors that need both real and fake examples. Their method uses a clever trick: by averaging multiple images, they highlight patterns that help tell real from fake. They then use a model to understand these patterns and decide if a single image is real or fake based on how well it fits. Tests show μFlow works well even on deepfakes it hasn’t seen before, beating previous detectors.
deepfake detectiongenerative adversarial networks (GANs)diffusion modelsone-class classificationnormalizing flowfeature representationout-of-distribution generalizationimage averaginglikelihood-based detection
Authors
Orazio Pontorno, Mattia Litrico, Luca Guarnera, Mario Valerio Giuffrida, Sebastiano Battiato
Abstract
Current generative models, including GANs and diffusion models, have reached an outstanding level of photorealism, posing significant risks to privacy and security. To ensure real-world applicability, deepfake detectors must generalise effectively to unseen generators. However, most existing approaches rely on supervised training with both real and fake images, which limits their generalisation especially across generators categories (e.g. GANs vs DMs). In this work, we introduce $μ$Flow, a one-class deepfake detector trained only on real images without relying on pseudo-deepfakes or synthetic artifacts. Our approach builds on the observation that averaging multiple images amplifies consistent generative traces, producing highly discriminative feature representations. We leverage this property by modelling the distribution of features extracted from averaged images and training a normalizing flow to align the feature space of individual images with this distribution. This alignment yields a likelihood-based criterion that separates real and fake samples while promoting strong generalisation. We evaluate $μ$Flow on a fully out-of-distribution setting, where both real and fake datasets are unseen during training. Experimental results show that our method significantly outperforms SOTA detectors. Project page: https://opontorno.github.io/MuFlow.