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

Despite huge success in the image domain, modern detection models such asFaster R-CNN have not been used nearly as much for video analysis. This isarguably due to the fact that detection models are designed to operate onsingle frames and as a result do not have a mechanism for learning motionrepresentations directly from video. We propose a learning procedure thatallows detection models such as Faster R-CNN to learn motion features directlyfrom the RGB video data while being optimized with respect to a pose estimationtask. Given a pair of video frames---Frame A and Frame B---we force our modelto predict human pose in Frame A using the features from Frame B. We do so byleveraging deformable convolutions across space and time. Our network learns tospatially sample features from Frame B in order to maximize pose detectionaccuracy in Frame A. This naturally encourages our network to learn motionoffsets encoding the spatial correspondences between the two frames. We referto these motion offsets as DiMoFs (Discriminative Motion Features).  In our experiments we show that our training scheme helps learn effectivemotion cues, which can be used to estimate and localize salient human motion.Furthermore, we demonstrate that as a byproduct, our model also learns featuresthat lead to improved pose detection in still-images, and better keypointtracking. Finally, we show how to leverage our learned model for the tasks ofspatiotemporal action localization and fine-grained action recognition.

Learning Discriminative Motion Features Through Detection