Title：Gesture recognition method and related equipment
Country：China
Patent No.：202010774742.5
Legal Status：Under review
Inventor：Xiangyu Li, Xiaomin Zheng
Assignee：Research Institute of Tsinghua University in Shenzhen
Address：Room A302, Research Institute of Tsinghua University in Shenzhen, the Southern District of the High-tech Industrial Park, Nanshan District, Shenzhen 518057, Guangdong
Filing Date：2020-08-04
Issue Date：
Abstract：
The invention provides a gesture recognition method and related equipment. The method comprises the steps of obtaining features of multiple standard gestures; training a binary classification random forest according to the characteristics of each standard gesture and forming a multi-classification random forest; acquiring features of a to-be-recognized gesture; judging the features of the to-be-recognized gesture for a preset number of rounds, and selecting a decision tree from each currently remaining binary classification random forest for judgment in each round of judgment; if the number ofthe currently remaining binary random forests is one, determining a standard gesture corresponding to the currently remaining binary random forests as a recognition result; if the decision tree of each binary random forest remaining at present is selected, determining the standard gesture corresponding to the binary random forest with the maximum positive conclusion vote number as a recognition result; and if the decision trees of the current remaining binary random forests are not selected completely, removing the binary random forests meeting pruning conditions from the multi-classificationrandom forests. The gesture recognition result can be quickly obtained.
Patent Certificate：
PDF/Jpg