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

In this paper, we tackle the problem of video alignment, the process ofmatching the frames of a pair of videos containing similar actions. The mainchallenge in video alignment is that accurate correspondence should beestablished despite the differences in the execution processes and appearancesbetween the two videos. We introduce an unsupervised method for alignment thatuses global and local features of the frames. In particular, we introduceeffective features for each video frame using three machine vision tools:person detection, pose estimation, and VGG network. Then, the features areprocessed and combined to construct a multidimensional time series thatrepresents the video. The resulting time series are used to align videos of thesame actions using a novel version of dynamic time warping named DiagonalizedDynamic Time Warping(DDTW). The main advantage of our approach is that notraining is required, which makes it applicable for any new type of actionwithout any need to collect training samples for it. For evaluation, weconsidered video synchronization and phase classification tasks on the Pennaction dataset. Also, for an effective evaluation of the video synchronizationtask, we present a new metric called Enclosed Area Error(EAE). The results showthat our method outperforms previous state-of-the-art methods, such as TCC, andother self-supervised and weakly supervised methods.

Video alignment using unsupervised learning of local and global features