Robust Image Steganography in Online Social Networks via Neural Style Transfer

Existing style-transfer steganography schemes suffer from three critical limitations: insufficient robustness against online social network (OSN) processing pipelines, susceptibility to steganalytic detection, and degraded visual quality. To address these challenges holistically, we propose StegTran...

Celý popis

Podrobná bibliografie

Hlavní autoři:	Peng Luo, Jia Liu, Qian Dang, Dejun Mu
Médium:	Artigo
Jazyk:	Inglês
Vydáno:	MDPI AG 2026-02-01
Edice:	Mathematics
Témata:	style-transfer steganography robustness adversarial training steganalysis visual quality distortion simulation
On-line přístup:	https://www.mdpi.com/2227-7390/14/4/629
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!