©iStockphoto.com/Jiraro praditcharoenkul Interaction with Three-Dimensional Gesture and Character Input in Virtual Reality Advanced Interaction and VR/AR By Jiachen Yang, Yafang Wang, Zhihan Lv, Na Jiang, and Anthony Steed Recognizing gestures in different directions and improving user input. H and-gesture recognition is a key aspect to make virtual-reality (VR) interaction more convenient. Making a user's idea understood by computers (i.e., character input) plays an important role in interaction. Current methods of hand gesture and character input are too limited to make full use of the powerful capacity possessed by current computers. In this article, we propose a natural three-dimensional (3-D) input method based on stereo cameras as an interface between human and machine. We segment the hand based on skin-color detection and train a neural network based on Hu moments to recognize valid and invalid gestures defined in our article. The Hu moments are seven invariant moments constituting a set of fea- Digital Object Identifier 10.1109/MCE.2017.2776500 Date of publication: 8 February 2018 64 IEEE Consumer Electronics Magazine ^ MARCH 2018 ©iStockphoto.com/erhui1979 ture vectors that are extracted and normalized after image preprocessing. A Hu moment is a rotation, scale, and translation invariant. For valid gestures, we implement stereo matching and 3-D coordinate calculation and line them up to formulate characters. Our method can robustly recognize 3-D gestures in different directions and make a user input freer when compared with traditional methods. 2162-2248/18©2018IEEEhttp://www.iStockphoto.com/Jiraro http://www.iStockphoto.com/erhui1979