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

Prevailing video frame interpolation algorithms, that generate theintermediate frames from consecutive inputs, typically rely on complex modelarchitectures with heavy parameters or large delay, hindering them from diversereal-time applications. In this work, we devise an efficient encoder-decoderbased network, termed IFRNet, for fast intermediate frame synthesizing. Itfirst extracts pyramid features from given inputs, and then refines thebilateral intermediate flow fields together with a powerful intermediatefeature until generating the desired output. The gradually refined intermediatefeature can not only facilitate intermediate flow estimation, but alsocompensate for contextual details, making IFRNet do not need additionalsynthesis or refinement module. To fully release its potential, we furtherpropose a novel task-oriented optical flow distillation loss to focus onlearning the useful teacher knowledge towards frame synthesizing. Meanwhile, anew geometry consistency regularization term is imposed on the graduallyrefined intermediate features to keep better structure layout. Experiments onvarious benchmarks demonstrate the excellent performance and fast inferencespeed of proposed approaches. Code is available athttps://github.com/ltkong218/IFRNet.

IFRNet: Intermediate Feature Refine Network for Efficient Frame Interpolation