Title :
Recognition of finger-pointing direction using color clustering and image segmentation
Author :
Asano, Hidetsugu ; Nagayasu, Takeshi ; Orimo, Tatsuya ; Terabayashi, Kenji ; Ohta, Mutsumi ; Umeda, Kazunori
Author_Institution :
R&D Division, PIONEER Corporation, Kanagawa, Japan
Abstract :
An accurate method of recognizing finger-pointing direction using multiple cameras is proposed. Input images are segmented by color clustering prior to shape analysis. Some restrictions in previous methods, such as narrowness of operation area and predefined skin color, are removed. Finger-pointing recognition in eight directions is achieved with an average of 90% recall and 93% precision.
Keywords :
Cameras; Estimation; Image color analysis; Image segmentation; Shape; Thumb; Gesture Interface; Human Machine Interface; Image Processing; Intelligent Room; Network Camera;
Conference_Titel :
SICE Annual Conference (SICE), 2013 Proceedings of
Conference_Location :
Nagoya, Japan
