A new multimodal, multiview, and interactive (M2I) dataset enables comprehensive evaluation of human action recognition across single-view, cross-view, cross-domain, and multitask learning scenarios. This dataset addresses limitations in existing benchmarks for robust algorithm comparison.
You might also read
Articles linked to this work by shared authors, journal, and citation graph.
Area of Science:
Background:
Purpose of the Study:
Main Methods:
Main Results:
Conclusions: