International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 176 - Number 14 |
Year of Publication: 2020 |
Authors: Aktar Hossen, Feng Xiufang, Xiong XiaoQiao, Zhang Xin |
10.5120/ijca2020920068 |
Aktar Hossen, Feng Xiufang, Xiong XiaoQiao, Zhang Xin . WiHRR: CSI based English Alphabet Handwriting Recognition. International Journal of Computer Applications. 176, 14 ( Apr 2020), 31-37. DOI=10.5120/ijca2020920068
Human Computer Interaction has been an important and vital subject of research since last few years and new and innovative techniques are being implemented in this field with each passing day. This research paper tries to improve existing samples of Channel frequency Response (CFR) that is obtained from Wi-Fi devices in the form of channel state information (CSI). In this research article, handwriting is used as an input but character recognition is not done on the paper but in the air.It discusses all the important procedures which are undertaken to accomplish this task which include fetching the input, pre-processing , feature extraction and classification of characters using machine learning algorithms.Current wireless devices can measure the interference caused by handwritten English alphabet in the range of signal coverage during the transmission of wireless signals in their own channels, and these measured interference s can be applied to handwritten recognition. Compared with traditional behavior recognition using cameras or wearable sensor devices, handwritten recognition using Wi-Fi wireless signals has more unexpected benefits: first, the wireless signals do not need LOS (line of sight) monitoring; second, no need to wear redundant sensor devices; and, at the same time, no need to wear redundant sensor devices. Wireless signals can be deployed in various scenarios without dead corners, such as home, office, public area and so on. Compared with the camera, the behavior recognition based on wireless signal is not affected by the weather.