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

Objective monitoring and assessment of human motor behavior can improve thediagnosis and management of several medical conditions. Over the past decade,significant advances have been made in the use of wearable technology forcontinuously monitoring human motor behavior in free-living conditions.However, wearable technology remains ill-suited for applications which requiremonitoring and interpretation of complex motor behaviors (e.g. involvinginteractions with the environment). Recent advances in computer vision and deeplearning have opened up new possibilities for extracting information from videorecordings. In this paper, we present a hierarchical vision-based behaviorphenotyping method for classification of basic human actions in videorecordings performed using a single RGB camera. Our method addresses challengesassociated with tracking multiple human actors and classification of actions invideos recorded in changing environments with different fields of view. Weimplement a cascaded pose tracker that uses temporal relationships betweendetections for short-term tracking and appearance-based tracklet fusion forlong-term tracking. Furthermore, for action classification, we use poseevolution maps derived from the cascaded pose tracker as low-dimensional andinterpretable representations of the movement sequences for training aconvolutional neural network. The cascaded pose tracker achieves an averageaccuracy of 88\% in tracking the target human actor in our video recordings,and overall system achieves average test accuracy of 84\% for target-specificaction classification in untrimmed video recordings.

Target-Specific Action Classification for Automated Assessment of Human Motor Behavior from Video