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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Automatic event detection from time series signals has wide applications,such as abnormal event detection in video surveillance and event detection ingeophysical data. Traditional detection methods detect events primarily by theuse of similarity and correlation in data. Those methods can be inefficient andyield low accuracy. In recent years, because of the significantly increasedcomputational power, machine learning techniques have revolutionized manyscience and engineering domains. In this study, we apply a deep-learning-basedmethod to the detection of events from time series seismic signals. However, adirect adaptation of the similar ideas from 2D object detection to our problemfaces two challenges. The first challenge is that the duration of earthquakeevent varies significantly; The other is that the proposals generated aretemporally correlated. To address these challenges, we propose a novel cascadedregion-based convolutional neural network to capture earthquake events indifferent sizes, while incorporating contextual information to enrich featuresfor each individual proposal. To achieve a better generalization performance,we use densely connected blocks as the backbone of our network. Because of thefact that some positive events are not correctly annotated, we furtherformulate the detection problem as a learning-from-noise problem. To verify theperformance of our detection methods, we employ our methods to seismic datagenerated from a bi-axial "earthquake machine" located at Rock MechanicsLaboratory, and we acquire labels with the help of experts. Through ournumerical tests, we show that our novel detection techniques yield highaccuracy. Therefore, our novel deep-learning-based detection methods canpotentially be powerful tools for locating events from time series data invarious applications.

Cascaded Region-based Densely Connected Network for Event Detection: A Seismic Application