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

Full face synthesis and partial face manipulation by virtue of the generativeadversarial networks (GANs) and its variants have raised wide public concerns.In the multi-media forensics area, detecting and ultimately locating the imageforgery has become an imperative task. In this work, we investigate thearchitecture of existing GAN-based face manipulation methods and observe thatthe imperfection of upsampling methods therewithin could be served as animportant asset for GAN-synthesized fake image detection and forgerylocalization. Based on this basic observation, we have proposed a novelapproach, termed FakeLocator, to obtain high localization accuracy, at fullresolution, on manipulated facial images. To the best of our knowledge, this isthe very first attempt to solve the GAN-based fake localization problem with agray-scale fakeness map that preserves more information of fake regions. Toimprove the universality of FakeLocator across multifarious facial attributes,we introduce an attention mechanism to guide the training of the model. Toimprove the universality of FakeLocator across different DeepFake methods, wepropose partial data augmentation and single sample clustering on the trainingimages. Experimental results on popular FaceForensics++, DFFD datasets andseven different state-of-the-art GAN-based face generation methods have shownthe effectiveness of our method. Compared with the baselines, our methodperforms better on various metrics. Moreover, the proposed method is robustagainst various real-world facial image degradations such as JPEG compression,low-resolution, noise, and blur.

FakeLocator: Robust Localization of GAN-Based Face Manipulations