English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

AIGC (AI-Generated Content) has achieved tremendous success in manyapplications such as text-to-image tasks, where the model can generatehigh-quality images with diverse prompts, namely, different descriptions innatural languages. More surprisingly, the emerging personalization techniqueseven succeed in describing unseen concepts with only a few personal images asreferences, and there have been some commercial platforms for sharing thevaluable personalized concept. However, such an advanced technique alsointroduces a severe threat, where malicious users can misuse the target conceptto generate highly-realistic illegal images. Therefore, it becomes necessaryfor the platform to trace malicious users and hold them accountable.  In this paper, we focus on guarding the most popular lightweightpersonalization model, ie, Textual Inversion (TI). To achieve it, we proposethe novel concept watermarking, where watermark information is embedded intothe target concept and then extracted from generated images based on thewatermarked concept. Specifically, we jointly train a watermark encoder and awatermark decoder with the sampler in the loop.  It shows great resilience to different diffusion sampling processes possiblychosen by malicious users, meanwhile preserving utility for normal use. Inpractice, the concept owner can upload his concept with different watermarks(ie, serial numbers) to the platform, and the platform allocates differentusers with different serial numbers for subsequent tracing and forensics.

Catch You Everything Everywhere: Guarding Textual Inversion via Concept Watermarking