Gesture Recognition - Key Insights and Future Directions

Explore key insights and future directions in gesture recognition through this keynote address delivered by Joseph J. LaViola at SUI 2019. Delve into the challenges of creating accurate and robust 3D gestural interfaces, examining research projects aimed at increasing recognizable gestures and evaluating ecological validity. Learn about pen- and touch-based interactive computing, 3D spatial interfaces, and human-robot interaction from LaViola's extensive experience. Gain valuable lessons and discover potential areas for future work in this complex field. The talk covers topics such as gesture segmentation, user-dependent and independent training, context-aware recognition, and in-game experiments, providing a comprehensive overview of the current state and future possibilities of gesture recognition technology.

Intro
Roadmap
Johnny Mnemonic
Andy Van Dam
What is a 3D gesture
Gesture segmentation
Nintendo Wiimote
OneManBand
Interactive Collaborative Music
Gesture Set
Gesture Collection
User Dependent Training
User Dependent Training Results
User Independent Training Results
Ecological Validity
Gesture Recognition in Context
ContextAware Gesture Recognition
Dynamic Time Warp Gesture Recognition
Linear classifier
Results
InGame Experiment
InGame Experiment Results
Jackknife
Correction Factors
Bounding Boxes
Evaluation
Parkour
Accuracy
Future Directions
Gestures in the Physical World
Robots
Knowledgeability
Moving Objects
Conclusion
Question