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

Deep learning based 3D stereo networks give superior performance compared to2D networks and conventional stereo methods. However, this improvement in theperformance comes at the cost of increased computational complexity, thusmaking these networks non-practical for the real-world applications.Specifically, these networks use 3D convolutions as a major work horse torefine and regress disparities. In this work first, we show that these 3Dconvolutions in stereo networks consume up to 94% of overall network operationsand act as a major bottleneck. Next, we propose a set of "plug-&-run" separableconvolutions to reduce the number of parameters and operations. When integratedwith the existing state of the art stereo networks, these convolutions lead upto 7x reduction in number of operations and up to 3.5x reduction in parameterswithout compromising their performance. In fact these convolutions lead toimprovement in their performance in the majority of cases.

Separable Convolutions for Optimizing 3D Stereo Networks