An Evaluation of Shape Changes for Conveying Emotions
summary
Two studies use a neoprene 2D flexible surface with 11 bend sensors and human-recorded deformations to map shapes to emotions; 20 creators produce common patterns (U for happiness, inverted-U for sadness, crumple for anger) and animated those into videos. Automated classification yields 45.8% precision and 40.4% recall, while observers label videos correctly 35% of the time with 12.3% “no emotion”; perceived valence/arousal align within ~28% of circumplex dimensions, showing viable but noisy emotion signaling.
AI-written summary · about
chapters (10)
- 0:00Introduction to study
- 1:54Background and motivation
- 3:43Shape parameters and frameworks
- 4:26First user study setup
- 6:11Emotion-shape correlations
- 8:46Machine learning classification
- 10:11Second user study with 3D models
- 13:28Results and key findings
- 14:33Conclusions and implications
- 15:57Q&A session
related talks
37mEmotion in Motion
53mKEYNOTE | The Power of Emotion and Video Marketing
24mDelightful UI Animations by Understanding the Brain
49mMotion & Playfulness
12mHeatFlow: A Thermal-Tactile Display for Dynamic 2D Thermal Movements
26mDelightful UI animations by understanding the brain
43mPurpose and Style: Using Motion with Meaning in Interfaces
24mMotional Intelligence: How Animation Brings Product to Life