Diffusion models meet image counter-forensics

Matías Tailanián, Marina Gardella, Alvaro Pardo, Pablo Musé

Producción científica: Capítulo del libro/informe/acta de congresoContribución a la conferenciarevisión exhaustiva

3 Citas (Scopus)

Resumen

From its acquisition in the camera sensors to its storage, different operations are performed to generate the final image. This pipeline imprints specific traces into the image to form a natural watermark. Tampering with an image disturbs these traces; these disruptions are clues that are used by most methods to detect and locate forgeries. In this article, we assess the capabilities of diffusion models to erase the traces left by forgers and, therefore, deceive forensics methods. Such an approach has been recently introduced for adversarial purification, achieving significant performance. We show that diffusion purification methods are well suited for counter-forensics tasks. Such approaches outperform already existing counter-forensics techniques both in deceiving forensics methods and in preserving the natural look of the purified images. The source code is publicly available at https://github.com/mtailanian/diff-cf.

Idioma originalInglés
Título de la publicación alojadaProceedings - 2024 IEEE Winter Conference on Applications of Computer Vision, WACV 2024
EditorialInstitute of Electrical and Electronics Engineers Inc.
Páginas3913-3923
Número de páginas11
ISBN (versión digital)9798350318920
DOI
EstadoPublicada - 3 ene. 2024
Evento2024 IEEE Winter Conference on Applications of Computer Vision, WACV 2024 - Waikoloa
Duración: 4 ene. 20248 ene. 2024

Serie de la publicación

NombreProceedings - 2024 IEEE Winter Conference on Applications of Computer Vision, WACV 2024

Conferencia

Conferencia2024 IEEE Winter Conference on Applications of Computer Vision, WACV 2024
País/TerritorioUnited States
CiudadWaikoloa
Período4/01/248/01/24

Huella

Profundice en los temas de investigación de 'Diffusion models meet image counter-forensics'. En conjunto forman una huella única.

Citar esto