Marozas Vaidotas, Janusauskas Artŭras, Lukosevicius Arŭnas, Sörnmo Leif
Department of Telecommunications and Electronics, Kaunas University of Technology, Lithuania.
IEEE Trans Biomed Eng. 2006 Aug;53(8):1586-93. doi: 10.1109/TBME.2006.876626.
This paper presents a unified approach to multiscale detection of transient evoked otoacoustic emissions (TEOAEs). Using statistical detection theory, it is shown that the optimal detector involves a time windowing operation where the window can be estimated from ensemble correlation information. The detector performs adaptive splitting of the signal into different frequency bands using either wavelet or wavelet packet decomposition. A simplified detector is proposed in which signal energy is omitted. The results show that the simplified detector performs significantly better than existing TEOAE detectors based on wave reproducibility or the modified variance ratio, whereas the detector involving signal energy does not offer such a performance advantage.
本文提出了一种用于瞬态诱发耳声发射（TEOAEs）多尺度检测的统一方法。利用统计检测理论表明，最优检测器涉及一个时间窗操作，其中窗口可根据总体相关信息进行估计。该检测器使用小波或小波包分解对信号进行自适应分解到不同频带。提出了一种简化的检测器，其中省略了信号能量。结果表明，简化后的检测器比现有的基于波重现性或修正方差比的TEOAEs检测器性能显著更好，而涉及信号能量的检测器并未提供这样的性能优势。
