Blood Cell Signal Recognition Based On Wavelet Packet Transform

Blood analyzer is essential commonly used for clinical blood routineexamination instrument，which hospital are widely used at home and abroad. Bloodanalyzer in improves the accuracy of test results also provides a number ofexperiments on the basis of the index, greatly improve the clinical diagnosis andtreatment effect and quality of service. In clinical trials, blood cell signal recognitionaccuracy and stability of the reflected the improvement of blood cell analyzer.For blood cell signal is non-stationary, non-Gaussian and nonlinear anduncertain characteristics, this topic by using the MATLAB software platform, usingwavelet packet transform and the method of combining the BP neural network foridentification of blood cells. Wavelet packet with a fully adaptability, hightime-frequency cohesion and the characteristics of strong noise resistance, is an idealsignal analysis method, especially suitable for the analysis of blood cell signalprocessing. Cells extracted using wavelet packet transform signal energy in space ortime scales on the distribution of characteristic vector and the input to the BP neuralnetwork training, finally to achieve the purpose of identifying the abnormal cells andnormal cells. Research based on wavelet packet transform blood signals for blood cellrecognition accuracy and is of significance to improve the accuracy, but also for theidentification and classification of the blood cells to explore a new way.