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/

Zendy Plus

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

Zendy Tools

Zendy Open

We propose a Few-shot Dynamic Neural Radiance Field (FDNeRF), the firstNeRF-based method capable of reconstruction and expression editing of 3D facesbased on a small number of dynamic images. Unlike existing dynamic NeRFs thatrequire dense images as input and can only be modeled for a single identity,our method enables face reconstruction across different persons with few-shotinputs. Compared to state-of-the-art few-shot NeRFs designed for modelingstatic scenes, the proposed FDNeRF accepts view-inconsistent dynamic inputs andsupports arbitrary facial expression editing, i.e., producing faces with novelexpressions beyond the input ones. To handle the inconsistencies betweendynamic inputs, we introduce a well-designed conditional feature warping (CFW)module to perform expression conditioned warping in 2D feature space, which isalso identity adaptive and 3D constrained. As a result, features of differentexpressions are transformed into the target ones. We then construct a radiancefield based on these view-consistent features and use volumetric rendering tosynthesize novel views of the modeled faces. Extensive experiments withquantitative and qualitative evaluation demonstrate that our method outperformsexisting dynamic and few-shot NeRFs on both 3D face reconstruction andexpression editing tasks. Code is available athttps://github.com/FDNeRF/FDNeRF.

FDNeRF: Few-shot Dynamic Neural Radiance Fields for Face Reconstruction and Expression Editing