Joints and their relations as critical features in action discrimination: Evidence from a classification image method
Feature (linguistics)
Similarity (geometry)
DOI:
10.1167/15.1.20
Publication Date:
2015-01-21T04:22:11Z
AUTHORS (2)
ABSTRACT
Classifying an action as a runner or walker is seemingly effortless process. However, it difficult to determine which features are used with hypothesis-driven research, because biological motion stimuli generally consist of about dozen joints, yielding enormous number potential relationships among them. Here, we develop hypothesis-free approach based on classification image method, using experimental data from relatively few trials (∼1,000 per subject). Employing ambiguous actions morphed between and runner, identified three types that play important roles in discriminating bipedal locomotion presented side view: (a) critical joint feature, supported by the finding similarity movements feet wrists prototypical these joints were most reliably across all participants; (b) structural features, indicated contributions almost other potentially through form-based analysis; (c) relational revealed statistical correlations contributions, specifically relations two feet, wrists/elbow hips. When inverted, only remained significantly influence discrimination responses. continuous depth rotation, associated strongly Using double-pass paradigm, estimated internal noise twice large external noise, consistent previous findings. Overall, our novel design rich set discrimination. The visual system flexibly selects subset depending viewing conditions.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (0)
CITATIONS (26)
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....