Improving Synthetically Generated Image Detection in Cross-Concept Settings

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F23%3A00371413" target="_blank" >RIV/68407700:21230/23:00371413 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1145/3592572.3592846" target="_blank" >https://doi.org/10.1145/3592572.3592846</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1145/3592572.3592846" target="_blank" >10.1145/3592572.3592846</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Improving Synthetically Generated Image Detection in Cross-Concept Settings

Popis výsledku v původním jazyce

New advancements for the detection of synthetic images are critical for fighting disinformation, as the capabilities of generative AI models continuously evolve and can lead to hyper-realistic synthetic imagery at unprecedented scale and speed. In this paper, we focus on the challenge of generalizing across different concept classes, e.g., when training a detector on human faces and testing on synthetic animal images - highlighting the ineffectiveness of existing approaches that randomly sample generated images to train their models. By contrast, we propose an approach based on the premise that the robustness of the detector can be enhanced by training it on realistic synthetic images that are selected based on their quality scores according to a probabilistic quality estimation model. We demonstrate the effectiveness of the proposed approach by conducting experiments with generated images from two seminal architectures, StyleGAN2 and Latent Diffusion, and using three different concepts for each, so as to measure the cross-concept generalization ability. Our results show that our quality-based sampling method leads to higher detection performance for nearly all concepts, improving the overall effectiveness of the synthetic image detectors.

Název v anglickém jazyce

Improving Synthetically Generated Image Detection in Cross-Concept Settings

Popis výsledku anglicky

New advancements for the detection of synthetic images are critical for fighting disinformation, as the capabilities of generative AI models continuously evolve and can lead to hyper-realistic synthetic imagery at unprecedented scale and speed. In this paper, we focus on the challenge of generalizing across different concept classes, e.g., when training a detector on human faces and testing on synthetic animal images - highlighting the ineffectiveness of existing approaches that randomly sample generated images to train their models. By contrast, we propose an approach based on the premise that the robustness of the detector can be enhanced by training it on realistic synthetic images that are selected based on their quality scores according to a probabilistic quality estimation model. We demonstrate the effectiveness of the proposed approach by conducting experiments with generated images from two seminal architectures, StyleGAN2 and Latent Diffusion, and using three different concepts for each, so as to measure the cross-concept generalization ability. Our results show that our quality-based sampling method leads to higher detection performance for nearly all concepts, improving the overall effectiveness of the synthetic image detectors.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

2nd ACM International Workshop on Multimedia AI against Disinformation (MAD '23)

ISBN

979-8-4007-0187-0

ISSN

—

e-ISSN

—

Počet stran výsledku

8

Strana od-do

28-35

Název nakladatele

ACM

Místo vydání

New York

Místo konání akce

Thessaloniki

Datum konání akce

12. 6. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001059176200005