摘要：

研究基于3D加速度传感器的空间手写识别技术，提出一种基于时频融合特征的分类识别方法。从加速度数据中提取短时能量 (STE)特征及低频分量，经快速傅里叶变换后提取频域特征WPD+FFT，将时域特征STE和频域特征WPD+FFT进行特征融合，利用主成分分析法对其降维，采用支持向量机进行分类识别。实验结果表明，该方法能提高空间手写识别系统的识别率。

关键词: 3D空间手写识别, 短时能量, 小波包分解, 快速傅里叶变换, 特征融合, 支持向量机

Abstract:

In the research of space handwriting recognition technology based on 3D accelerometer, a recognition method based on time-frequency fusion feature is proposed. From accelerometer data, it extracts the Short-time Energy(STE) feature. The hybrid feature which combines Wavelet Packet Decomposition with Fast Fourier Transform(WPD+FFT) are extracted, then the above two categories features are fused together and the Principal Component Analysis(PCA) is employed to reduce the dimension of the fusion feature. Supported Vector Machine(SVM) is used in recognition. Experimental results show that the proposed method can improve the performance of 3D space handwriting recognition system.

Key words: 3D space handwriting recognition, Short-time Energy(STE), Wavelet Packet Decomposition(WPD), Fast Fourier Transform(FFT), feature fusion, Support Vector Machine(SVM)

中图分类号: 

• TP212.9