The feature extraction of signal content into semantic parts plays a key role in applications such as analysis, enhancement, compression, restoration, and more, is a foundational and key technique for many fields such as pattern recognition , intelligent system and machinery fault diagnosis. Although proposed many approaches to tackle this problem in recent years, they had many disadvantages, Nonparametric basis feature extraction ( NBFE ) is a decomposition method based on sparse representation of signals and images. The main idea of NBFE is to decompose the observed signal into a series expansions of waveforms which best matches the signal local structures. These waveforms are calculated by the nonparametric waveform estimation method. This paper introduced the theory of NBFE. Also, it described the advances on NBFE. Finally, this paper pointed out the several main problems and anticipated further research directions.%有效的信号特征提取技术在信号的分析、增强、压缩、复原等领域起着重要的作用,是模式识别、智能系统和故障诊断等诸多领域的基础和关键.虽然目前研究者提出了很多方法来解决这个问题,然而处理效果并不理想.非参数基函数特征提取是一种基于稀疏表示的特征提取方法,方法的核心是将观察信号分解为一组最好匹配信号局部结构的特征波形的线性展开,这些特征波形是由非参数基函数特征波形估计方法计算所得.详细描述了非参数基函数特征提取方法的理论思想,介绍了该方法的最新研究进展及其存在的问题,最后指出了该方法进一步发展的方向.
展开▼
