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

We present DietNeRF, a 3D neural scene representation estimated from a fewimages. Neural Radiance Fields (NeRF) learn a continuous volumetricrepresentation of a scene through multi-view consistency, and can be renderedfrom novel viewpoints by ray casting. While NeRF has an impressive ability toreconstruct geometry and fine details given many images, up to 100 forchallenging 360{\deg} scenes, it often finds a degenerate solution to its imagereconstruction objective when only a few input views are available. To improvefew-shot quality, we propose DietNeRF. We introduce an auxiliary semanticconsistency loss that encourages realistic renderings at novel poses. DietNeRFis trained on individual scenes to (1) correctly render given input views fromthe same pose, and (2) match high-level semantic attributes across different,random poses. Our semantic loss allows us to supervise DietNeRF from arbitraryposes. We extract these semantics using a pre-trained visual encoder such asCLIP, a Vision Transformer trained on hundreds of millions of diversesingle-view, 2D photographs mined from the web with natural languagesupervision. In experiments, DietNeRF improves the perceptual quality offew-shot view synthesis when learned from scratch, can render novel views withas few as one observed image when pre-trained on a multi-view dataset, andproduces plausible completions of completely unobserved regions.

Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis