Online Continuous Scale Estimation of Hand Gestures

Woosuk KIM  Hideaki KUZUOKA  Kenji SUZUKI  

IEICE TRANSACTIONS on Information and Systems   Vol.E95-D   No.10   pp.2447-2455
Publication Date: 2012/10/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.E95.D.2447
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Human-computer Interaction
hand gestures,  gesture synchronization,  scale estimation,  longest common subsequence (LCSS),  

Full Text: PDF>>
Buy this Article

The style of a gesture provides significant information for communication, and thus understanding the style is of great importance in improving gestural interfaces using hand gestures. We present a novel method to estimate temporal and spatial scale—which are considered principal elements of the style—of hand gestures. Gesture synchronization is proposed for matching progression between spatio-temporally varying gestures, and scales are estimated based on the progression matching. For comparing gestures of various sizes and speeds, gesture representation is defined by adopting turning angle representation. Also, LCSS is used as a similarity measure for reliability and robustness to noise and outliers. Performance of our algorithm is evaluated with synthesized data to show the accuracy and robustness to noise and experiments are carried out using recorded hand gestures to analyze applicability under real-world situations.