Enhancing Abnormality Identification: Robust Out-of-Distribution Strategies for Deepfake Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Maiano, Luca, Casadei, Fabrizio, Amerini, Irene
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912411500412928
author Maiano, Luca
Casadei, Fabrizio
Amerini, Irene
author_facet Maiano, Luca
Casadei, Fabrizio
Amerini, Irene
contents Detecting deepfakes has become a critical challenge in Computer Vision and Artificial Intelligence. Despite significant progress in detection techniques, generalizing them to open-set scenarios continues to be a persistent difficulty. Neural networks are often trained on the closed-world assumption, but with new generative models constantly evolving, it is inevitable to encounter data generated by models that are not part of the training distribution. To address these challenges, in this paper, we propose two novel Out-Of-Distribution (OOD) detection approaches. The first approach is trained to reconstruct the input image, while the second incorporates an attention mechanism for detecting OODs. Our experiments validate the effectiveness of the proposed approaches compared to existing state-of-the-art techniques. Our method achieves promising results in deepfake detection and ranks among the top-performing configurations on the benchmark, demonstrating their potential for robust, adaptable solutions in dynamic, real-world applications.
format Preprint
id arxiv_https___arxiv_org_abs_2506_02857
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Abnormality Identification: Robust Out-of-Distribution Strategies for Deepfake Detection
Maiano, Luca
Casadei, Fabrizio
Amerini, Irene
Computer Vision and Pattern Recognition
Detecting deepfakes has become a critical challenge in Computer Vision and Artificial Intelligence. Despite significant progress in detection techniques, generalizing them to open-set scenarios continues to be a persistent difficulty. Neural networks are often trained on the closed-world assumption, but with new generative models constantly evolving, it is inevitable to encounter data generated by models that are not part of the training distribution. To address these challenges, in this paper, we propose two novel Out-Of-Distribution (OOD) detection approaches. The first approach is trained to reconstruct the input image, while the second incorporates an attention mechanism for detecting OODs. Our experiments validate the effectiveness of the proposed approaches compared to existing state-of-the-art techniques. Our method achieves promising results in deepfake detection and ranks among the top-performing configurations on the benchmark, demonstrating their potential for robust, adaptable solutions in dynamic, real-world applications.
title Enhancing Abnormality Identification: Robust Out-of-Distribution Strategies for Deepfake Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.02857