A Real-Time Hand Gesture Recognition Approach Based on Motion Features of Feature Points

Dynamic hand gesture recognition enables people to communicate with computers naturally without any mechanical devices. Due to the spread of depth sensor such as Microsoft Kinect and Leap Motion, dynamic hand gesture recognition becomes possible for recognizing meticulous gesture information in real...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2014 IEEE 17th International Conference on Computational Science and Engineering S. 1096 - 1102
Hauptverfasser: Yingying She, Qian Wang, Yunzhe Jia, Ting Gu, Qun He, Baorong Yang
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.12.2014
Schlagworte:
Classification algorithms
Feature extraction
feature points
Gesture recognition
hand gesture recognition
HCI
Joints
Labeling
motion feature
Thumb
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Dynamic hand gesture recognition enables people to communicate with computers naturally without any mechanical devices. Due to the spread of depth sensor such as Microsoft Kinect and Leap Motion, dynamic hand gesture recognition becomes possible for recognizing meticulous gesture information in real time. However, most of these methods recognize the hand gesture by fuzzy features such as contour size, which cause imprecise hand gesture recognition. This paper presents a precise tracing of feature points including palm center, fingertips and joints by using Kinect. A novel recognition method based on precise motion features of these feature points is also presented. Having been tested with a series of applications, our method is proved to be robust and effective, and suitable for further application in real-time HCI systems.
DOI:10.1109/CSE.2014.216