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

Super-resolution (SR) techniques have recently been proposed to upscale theoutputs of neural radiance fields (NeRF) and generate high-quality images withenhanced inference speeds. However, existing NeRF+SR methods increase trainingoverhead by using extra input features, loss functions, and/or expensivetraining procedures such as knowledge distillation. In this paper, we aim toleverage SR for efficiency gains without costly training or architecturalchanges. Specifically, we build a simple NeRF+SR pipeline that directlycombines existing modules, and we propose a lightweight augmentation technique,random patch sampling, for training. Compared to existing NeRF+SR methods, ourpipeline mitigates the SR computing overhead and can be trained up to 23xfaster, making it feasible to run on consumer devices such as the AppleMacBook. Experiments show our pipeline can upscale NeRF outputs by 2-4x whilemaintaining high quality, increasing inference speeds by up to 18x on an NVIDIAV100 GPU and 12.8x on an M1 Pro chip. We conclude that SR can be a simple buteffective technique for improving the efficiency of NeRF models for consumerdevices.

FastSR-NeRF: Improving NeRF Efficiency on Consumer Devices with A Simple Super-Resolution Pipeline