TY - JOUR
T1 - A virtual mouse interface with a two-layered Bayesian network
AU - Roh, Myung Cheol
AU - Kang, Dongoh
AU - Huh, Sungju
AU - Lee, Seong Whan
N1 - Funding Information:
This work was partly supported by the ICT R&D program of MSIP/IITP [B0101-15-0552 , Development of Predictive Visual Intelligence Technology] and also supported by the Implementation of Technologies for Identification, Behavior, and Location of Human based on Sensor Network Fusion Program through the Ministry of Trade, Industry and Energy (Grant No. 10041629).
Publisher Copyright:
© 2015, Springer Science+Business Media New York.
PY - 2017/1/1
Y1 - 2017/1/1
N2 - During the last decade, many natural interaction methods between human and computer have been introduced. They were developed for substitutions of keyboard and mouse devices so that they provide convenient interfaces. Recently, many studies on vision based gestural control methods for Human-Computer Interaction (HCI) have been attracted attention because of their convenience and simpleness. Two of the key issues in these kinds of interfaces are robustness and real-time processing. This paper presents a hand gesture based virtual mouse interface and Two-layer Bayesian Network (TBN) for robust hand gesture recognition in real-time. The TBN provides an efficient framework to infer hand postures and gestures not only from information at the current time frame, but also from the preceding and following information, so that it compensates for erroneous postures and its locations under cluttered background environment. Experiments demonstrated that the proposed model recognized hand gestures with a recognition rate of 93.76 % and 85.15 % on simple and cluttered background video data, respectively, and outperformed previous methods: Hidden Markov Model (HMM), Finite State Machine (FSM).
AB - During the last decade, many natural interaction methods between human and computer have been introduced. They were developed for substitutions of keyboard and mouse devices so that they provide convenient interfaces. Recently, many studies on vision based gestural control methods for Human-Computer Interaction (HCI) have been attracted attention because of their convenience and simpleness. Two of the key issues in these kinds of interfaces are robustness and real-time processing. This paper presents a hand gesture based virtual mouse interface and Two-layer Bayesian Network (TBN) for robust hand gesture recognition in real-time. The TBN provides an efficient framework to infer hand postures and gestures not only from information at the current time frame, but also from the preceding and following information, so that it compensates for erroneous postures and its locations under cluttered background environment. Experiments demonstrated that the proposed model recognized hand gestures with a recognition rate of 93.76 % and 85.15 % on simple and cluttered background video data, respectively, and outperformed previous methods: Hidden Markov Model (HMM), Finite State Machine (FSM).
KW - Hand gesture recognition
KW - Two-layer Bayesian network
KW - Virtual mouse interface
UR - http://www.scopus.com/inward/record.url?scp=85009511130&partnerID=8YFLogxK
U2 - 10.1007/s11042-015-3144-x
DO - 10.1007/s11042-015-3144-x
M3 - Article
AN - SCOPUS:85009511130
SN - 1380-7501
VL - 76
SP - 1615
EP - 1638
JO - Multimedia Tools and Applications
JF - Multimedia Tools and Applications
IS - 2
ER -