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

A critical limitation of current methods based on Neural Radiance Fields(NeRF) is that they are unable to quantify the uncertainty associated with thelearned appearance and geometry of the scene. This information is paramount inreal applications such as medical diagnosis or autonomous driving where, toreduce potentially catastrophic failures, the confidence on the model outputsmust be included into the decision-making process. In this context, weintroduce Conditional-Flow NeRF (CF-NeRF), a novel probabilistic framework toincorporate uncertainty quantification into NeRF-based approaches. For thispurpose, our method learns a distribution over all possible radiance fieldsmodelling which is used to quantify the uncertainty associated with themodelled scene. In contrast to previous approaches enforcing strong constraintsover the radiance field distribution, CF-NeRF learns it in a flexible and fullydata-driven manner by coupling Latent Variable Modelling and ConditionalNormalizing Flows. This strategy allows to obtain reliable uncertaintyestimation while preserving model expressivity. Compared to previousstate-of-the-art methods proposed for uncertainty quantification in NeRF, ourexperiments show that the proposed method achieves significantly lowerprediction errors and more reliable uncertainty values for synthetic novel viewand depth-map estimation.

Conditional-Flow NeRF: Accurate 3D Modelling with Reliable Uncertainty Quantification