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

Recently, self-supervised learning has proved to be effective to learnrepresentations of events suitable for temporal segmentation in imagesequences, where events are understood as sets of temporally adjacent imagesthat are semantically perceived as a whole. However, although this approachdoes not require expensive manual annotations, it is data hungry and suffersfrom domain adaptation problems. As an alternative, in this work, we propose anovel approach for learning event representations named Dynamic Graph Embedding(DGE). The assumption underlying our model is that a sequence of images can berepresented by a graph that encodes both semantic and temporal similarity. Thekey novelty of DGE is to learn jointly the graph and its graph embedding. Atits core, DGE works by iterating over two steps: 1) updating the graphrepresenting the semantic and temporal similarity of the data based on thecurrent data representation, and 2) updating the data representation to takeinto account the current data graph structure. The main advantage of DGE overstate-of-the-art self-supervised approaches is that it does not require anytraining set, but instead learns iteratively from the data itself alow-dimensional embedding that reflects their temporal and semantic similarity.Experimental results on two benchmark datasets of real image sequences capturedat regular time intervals demonstrate that the proposed DGE leads to eventrepresentations effective for temporal segmentation. In particular, it achievesrobust temporal segmentation on the EDUBSeg and EDUBSeg-Desc benchmarkdatasets, outperforming the state of the art. Additional experiments on twoHuman Motion Segmentation benchmark datasets demonstrate the generalizationcapabilities of the proposed DGE.

Learning event representations for temporal segmentation of image sequences by dynamic graph embedding