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

The COVID-19 pandemic raises the problem of adapting face recognition systemsto the new reality, where people may wear surgical masks to cover their nosesand mouths. Traditional data sets (e.g., CelebA, CASIA-WebFace) used fortraining these systems were released before the pandemic, so they now seemunsuited due to the lack of examples of people wearing masks. We propose amethod for enhancing data sets containing faces without masks by creatingsynthetic masks and overlaying them on faces in the original images. Our methodrelies on SparkAR Studio, a developer program made by Facebook that is used tocreate Instagram face filters. In our approach, we use 9 masks of differentcolors, shapes and fabrics. We employ our method to generate a number of445,446 (90%) samples of masks for the CASIA-WebFace data set and 196,254(96.8%) masks for the CelebA data set, releasing the mask images athttps://github.com/securifai/masked_faces. We show that our method producessignificantly more realistic training examples of masks overlaid on faces byasking volunteers to qualitatively compare it to other methods or data setsdesigned for the same task. We also demonstrate the usefulness of our method byevaluating state-of-the-art face recognition systems (FaceNet, VGG-face,ArcFace) trained on our enhanced data sets and showing that they outperformequivalent systems trained on original data sets (containing faces withoutmasks) or competing data sets (containing masks generated by related methods),when the test benchmarks contain masked faces.

A realistic approach to generate masked faces applied on two novel masked face recognition data sets