Dual-domain self-supervised feature alignment via spectral–spatial representation learning for deepfake anomaly detection
- 1School of Computer and Data Science, Henan University of Urban Construction, Pingdingshan, Henan, China
Abstract
Deepfake anomaly detection has become increasingly critical in visual security applications. However, many existing methods rely heavily on large-scale supervised forgery annotations and often struggle to capture subtle manipulation traces distributed across both spatial and frequency domains. To address these limitations, we propose a Self-Supervised Dual-Domain Alignment (SDDA) framework that jointly learns spatial and spectral representations without requiring labeled forged samples. Specifically, SDDA employs two parameter-sharing encoders to extract complementary features from RGB images and their frequency-domain counterparts, while a cross-domain alignment module enforces consistency between spatial textures and spectral signatures. To provide explicit and reproducible self-supervised supervision, four complementary descriptors are further derived from FAN feature maps: Local Structural Deviation (LSD) for local gradient inconsistency, Global Pattern Discrepancy (GPD) for holistic channel-correlation differences, Local Residual Difference (LRD) for residual-level inconsistency, and Total Consistency Deviation (TCD) for multi-layer feature deviation. These descriptors are predicted from the fused spatial–frequency embedding, encouraging the model to learn manipulation-sensitive representations in the absence of fake samples. In addition, a dual-domain contrastive objective enhances the discrimination of subtle forgery-related anomalies while improving robustness against real-world degradations such as compression and illumination variations. Experimental results demonstrate that SDDA consistently outperforms state-of-the-art baselines, particularly under challenging unseen-manipulation scenarios. Overall, the proposed framework provides an interpretable and label-efficient solution for dual-domain deepfake anomaly detection.
Related
- Semi-supervised clustering with knowledge-guided representation learning in cryo-electron tomography
- Comparison of anomaly detection and supervised learning for acute myeloid leukemia blast identification: A proof-of-concept study
- A representation-level wavelet–EMD fusion framework for spectral analysis and detection of epileptiform patterns in pediatric EEG: computational validation with a roadmap for external benchmarking
- Enhancing Clinical Classifier Robustness with Deep Representation Learning Based OOD Detection on DNA Methylation Data
- Learn the interactions: Weakly supervised video anomaly detection with human-object interactions
BibTeX
@article{he2026dualdomain,
title = {{Dual-domain self-supervised feature alignment via spectral–spatial representation learning for deepfake anomaly detection}},
author = {Yi He},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0358049},
url = {https://doi.org/10.1371/journal.pone.0358049},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Dual-domain self-supervised feature alignment via spectral–spatial representation learning for deepfake anomaly detection
AU - Yi He
PY - 2026
DA - 2026/09/11
JO - PLOS One
DO - 10.1371/journal.pone.0358049
UR - https://doi.org/10.1371/journal.pone.0358049
AB - Deepfake anomaly detection has become increasingly critical in visual security applications. However, many existing methods rely heavily on large-scale supervised forgery annotations and often struggle to capture subtle manipulation traces distributed across both spatial and frequency domains. To address these limitations, we propose a Self-Supervised Dual-Domain Alignment (SDDA) framework that jointly learns spatial and spectral representations without requiring labeled forged samples. Specifically, SDDA employs two parameter-sharing encoders to extract complementary features from RGB images and their frequency-domain counterparts, while a cross-domain alignment module enforces consistency between spatial textures and spectral signatures. To provide explicit and reproducible self-supervised supervision, four complementary descriptors are further derived from FAN feature maps: Local Structural Deviation (LSD) for local gradient inconsistency, Global Pattern Discrepancy (GPD) for holistic channel-correlation differences, Local Residual Difference (LRD) for residual-level inconsistency, and Total Consistency Deviation (TCD) for multi-layer feature deviation. These descriptors are predicted from the fused spatial–frequency embedding, encouraging the model to learn manipulation-sensitive representations in the absence of fake samples. In addition, a dual-domain contrastive objective enhances the discrimination of subtle forgery-related anomalies while improving robustness against real-world degradations such as compression and illumination variations. Experimental results demonstrate that SDDA consistently outperforms state-of-the-art baselines, particularly under challenging unseen-manipulation scenarios. Overall, the proposed framework provides an interpretable and label-efficient solution for dual-domain deepfake anomaly detection.
SN - 1932-6203
ER -