Handwritten Character Recognition Using Orientation Quantization Based on 3D Accelerometer

Shiqi Zhang; Chun Yuan; Yan Zhang

1st International ICST Workshop on Human Control of Ubiquitous Systems

Research Article

Handwritten Character Recognition Using Orientation Quantization Based on 3D Accelerometer

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.4108/ICST.MOBIQUITOUS2008.3898,
    author={Shiqi Zhang and Chun Yuan and Yan Zhang},
    title={Handwritten Character Recognition Using Orientation Quantization Based on 3D Accelerometer},
    proceedings={1st International ICST Workshop on Human Control of Ubiquitous Systems},
    publisher={ACM},
    proceedings_a={HUCUBIS},
    year={2010},
    month={5},
    keywords={Handwritten Character Recognition Accelerometer Hidden Markov Model Quantization based on Orientation MEMS},
    doi={10.4108/ICST.MOBIQUITOUS2008.3898}
}

Shiqi Zhang
Chun Yuan
Yan Zhang
Year: 2010
Handwritten Character Recognition Using Orientation Quantization Based on 3D Accelerometer
HUCUBIS
ICST
DOI: 10.4108/ICST.MOBIQUITOUS2008.3898

Shiqi Zhang^1,2^,*, Chun Yuan¹^,*, Yan Zhang²^,*

1: Information Sci. and Tec. Division, Graduate School at Shenzhen, Tsinghua University Nanshan, Shenzhen, Guangdong, China
2: Dept. of Electronic and Info. Eng., Shenzhen Graduate School, Harbin Institute of Technology, Nanshan, Shenzhen, Guangdong, China

*Contact email: zhangshiqihit@yahoo.com.cn, yuanc@sz.tsinghua.edu.cn, ianzhang@hit.edu.cn

Abstract

This paper presents an online handwritten character recognition system. The whole system includes three parts: acceleration signal detection, signal processing and recognition by Hidden Markov Model (HMM). In hardware aspect, a mini-board with a three-dimensional accelerometer and a microcontroller is used to get real time acceleration values and send them to a terminal continuously. After effective section extraction and lowpass filtering, different quantizing methods based on acceleration orientation are used to quantize numerous data into small integral vectors. At last, we use HMM to do the recognition. For the experiments with 10 Arabic numerals, this system shows a high Recognition Rate (R.R.) of 94.29% in the database of 42 models for every Arabic numeral. This system could be used to reduce the size of handheld devices by discarding number keys and make human computer interaction more convenient and interesting.

Keywords: Handwritten Character Recognition Accelerometer Hidden Markov Model Quantization based on Orientation MEMS

Published: 2010-05-16
Publisher: ACM
Modified: 2010-05-16

: http://dx.doi.org/10.4108/ICST.MOBIQUITOUS2008.3898