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

Convolutional Neural Networks (CNNs) have been regarded as the go-to modelsfor visual recognition. More recently, convolution-free networks, based onmulti-head self-attention (MSA) or multi-layer perceptrons (MLPs), become moreand more popular. Nevertheless, it is not trivial when utilizing thesenewly-minted networks for video recognition due to the large variations andcomplexities in video data. In this paper, we present MLP-3D networks, a novelMLP-like 3D architecture for video recognition. Specifically, the architectureconsists of MLP-3D blocks, where each block contains one MLP applied acrosstokens (i.e., token-mixing MLP) and one MLP applied independently to each token(i.e., channel MLP). By deriving the novel grouped time mixing (GTM)operations, we equip the basic token-mixing MLP with the ability of temporalmodeling. GTM divides the input tokens into several temporal groups andlinearly maps the tokens in each group with the shared projection matrix.Furthermore, we devise several variants of GTM with different groupingstrategies, and compose each variant in different blocks of MLP-3D network bygreedy architecture search. Without the dependence on convolutions or attentionmechanisms, our MLP-3D networks achieves 68.5\%/81.4\% top-1 accuracy onSomething-Something V2 and Kinetics-400 datasets, respectively. Despite withfewer computations, the results are comparable to state-of-the-art widely-used3D CNNs and video transformers. Source code is available athttps://github.com/ZhaofanQiu/MLP-3D.

MLP-3D: A MLP-like 3D Architecture with Grouped Time Mixing