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

Current state-of-the-art models for video action recognition are mostly basedon expensive 3D ConvNets. This results in a need for large GPU clusters totrain and evaluate such architectures. To address this problem, we present alightweight and memory-friendly architecture for action recognition thatperforms on par with or better than current architectures by using only afraction of resources. The proposed architecture is based on a combination of adeep subnet operating on low-resolution frames with a compact subnet operatingon high-resolution frames, allowing for high efficiency and accuracy at thesame time. We demonstrate that our approach achieves a reduction by $3\sim4$times in FLOPs and $\sim2$ times in memory usage compared to the baseline. Thisenables training deeper models with more input frames under the samecomputational budget. To further obviate the need for large-scale 3Dconvolutions, a temporal aggregation module is proposed to model temporaldependencies in a video at very small additional computational costs. Ourmodels achieve strong performance on several action recognition benchmarksincluding Kinetics, Something-Something and Moments-in-time. The code andmodels are available at https://github.com/IBM/bLVNet-TAM.

More Is Less: Learning Efficient Video Representations by Big-Little Network and Depthwise Temporal Aggregation