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

The goal of this paper is to self-train a 3D convolutional neural network onan unlabeled video collection for deployment on small-scale video collections.As smaller video datasets benefit more from motion than appearance, we striveto train our network using optical flow, but avoid its computation duringinference. We propose the first motion-augmented self-training regime, we callMotionFit. We start with supervised training of a motion model on a small, andlabeled, video collection. With the motion model we generate pseudo-labels fora large unlabeled video collection, which enables us to transfer knowledge bylearning to predict these pseudo-labels with an appearance model. Moreover, weintroduce a multi-clip loss as a simple yet efficient way to improve thequality of the pseudo-labeling, even without additional auxiliary tasks. Wealso take into consideration the temporal granularity of videos duringself-training of the appearance model, which was missed in previous works. As aresult we obtain a strong motion-augmented representation model suited forvideo downstream tasks like action recognition and clip retrieval. Onsmall-scale video datasets, MotionFit outperforms alternatives for knowledgetransfer by 5%-8%, video-only self-supervision by 1%-7% and semi-supervisedlearning by 9%-18% using the same amount of class labels.

Motion-Augmented Self-Training for Video Recognition at Smaller Scale