微笑表情的近红外额耳穴信号小波超低频功率谱特征提取

Smiling Expression Feature Extraction Based on Wavelet Super-Slow Power Spectrum of Near Infrared Prefrontal Cortex Otopoint Signal

提出一种抽取微笑表情特征的近红外额耳穴信号小波分析方法,优点是无损间接测量脑额区的神经递质涨落。针对分别呈现平静或微笑表情的16名大学生被测者(男11人,女5人,平均年龄24±4岁),该方法选取右耳额穴,采集960nm近红外透射信号,做10级小波分解与重构,构造出超低频功率谱特征;统计结果表明,81%的受试者具有微笑小于平静的小波超低频功率谱特征。

A novel feature extraction method for smiling expression is proposed based on wavelet super-slow power spectrum of near infrared prefrontal cortex otopoint signal. The method merits in non-invasive indirect measurement of neurotransmitters fluctuation at subject＇s prefrontal cortex. Firstly, n = 16 subjects （5 females and 11 males） aged （24 ± 4） years old are selected showing separately in quiet or smiling expressions. And 960 nm near infrared transmissive signals at right-side prefrontal cortex otopoint are sampled in 1 000 Hz with total data length of 10 000- dot and filtered in band-pass of 0.1 Hz -40 Hz. Then, wavelet analysis with 10-stage decomposition and recon- struction is made for extracting the super-slow power spectrum feature of quiet versus smiling expression. Finally, the statistic result showed that 81% of subjects in smiling are with smaller feature values than in quiet expression.