Research On Multifocal Visual Electrophysiology Signal Detection And Processing

Multifocal visual electrophysiology detection technology is an important new progress in the field of visual electrophysiology, which can simultaneously detect ERG and visual evoked potential of hundreds of local areas in the short period of time and calculate the first and second order partial response by applying advanced computer and digital signal processing technology. Functions of each part of the retina can be understood by the respective reaction parameters. Multifocal visual electrophysiology detection technology is a new theory and technology used in clinical ophthalmology in recent years. The multifocal visual electrophysiology signal detection system is designed under the basic principle of multifocal visual electrophysiology. Since visual electrophysiology signals produced by the body are very weak and contains large background noise interferences, it's difficult to extract visual electrophysiology signals. Removal of the noise is the first part of multifocal visual electrophysiological signals data processing. Wavelet transform has a good localization property and multi-resolution analysis function in both time and frequency domain and can separate signals and noise, therefore, wavelet transform can be applied to visual electrophysiological signals and noise separation, improve signals' quality and acquire signals satisfying the requirements of clinical examination. Around this problem, the paper's research agenda focuses on four areas as follows:First, in order to detect multifocal visual electrophysiological signals, the multifocal visual electrophysiology detection system is completed according to basic principles of electrophysiology. The system design includes stimulate the system design, amplifier design, the extraction and separation of signal using system identification theory and the rapid separation method which extracts the impulse response of retinal local area from the total retinal response signal using fast Walsh-Hadamard transform. Multifocal visual electrophysiological signals are detected with the system design.Secondly, the wavelet transform theory is studied, the definition and principle of wavelet transform is summarized and the multi-resolution and fast wavelet transform algorithm-Mallat algorithm is analyzed. Principles and advantages of several de-noising methods such as the wavelet decomposition and reconstruction method, the modulus maxima de-noising method, the spatial correlation method, the translation invariant method the and wavelet threshold de-noising method are analyzed and compared, problems of decomposition levels in the wavelet de-noising process, threshold selection criteria and the selection of threshold function is discussed.Thirdly, wavelet decomposition level is determined, the wavelet function and wavelet threshold value is selected through analyzing and comparing the signal to noise ratio, root mean square error, energy ratio. Different methods of wavelet de-noising are used to de-noise the total retinal response signal, de-noising effects are compared, and the comparison chart of total response signal before and after de-noising is given.At last, the multifocal visual electrophysiology inspection system processes the total response signal and generates mfERG and mfVEP waveforms of various small regions, compared with unprocessed ones. Simulation results show that the wavelet de-noising method studied in this paper can accurately extract the waveform of local response signal, which meets the requirements of clinical medicine on multifocal visual electrophysiology response testing, analysis and diagnostic.Results of this paper provide some reference value to multifocal visual electrophysiology testing technology development and applications, and have high practical significance for clinical medicine.