GestureLens: Visual Analysis of Gestures in Presentation Videos

Area of Science:

• Human-Computer Interaction
• Communication Studies
• Data Visualization

Background:

• Effective gestures enhance public speaking and audience engagement.
• Manual analysis of gestures in presentation videos is time-consuming and inefficient.
• Existing methods lack efficient tools for exploring temporal gesture evolution and speech correlation.

Purpose of the Study:

• To introduce GestureLens, a visual analytics system for analyzing gesture usage in public speaking videos.
• To facilitate efficient gesture exploration and analysis for public speaking coaches.
• To enhance gesture training by correlating hand movements with speech content.

Main Methods:

• Developed GestureLens, a visual analytics system with exploration, relation, video, and dynamic views.
• Utilized heatmap aggregation for spatial gesture patterns and timeline decomposition for temporal patterns.
• Implemented linked analysis and intuitive glyph designs for exploring speech-gesture correlations.

Main Results:

• GestureLens provides an overview of gesture spatial and temporal distributions.
• The system reveals spatial patterns through gesture space heatmaps.
• It uncovers temporal patterns by decomposing hand movements into timelines and links gestures to speech content.

Conclusions:

• GestureLens effectively facilitates gesture exploration and analysis in presentation videos.
• The system's visual analytic approach improves gesture training for public speaking coaches.
• GestureLens demonstrates usefulness in understanding the complex relationship between gestures and speech.