Research On Noise Reduction Algorithm And Fabrication Improvement Of CI Microelectrode

Currently,the electronic cochlea is the only device in the world that can help some hearing-impaired deaf patients recovering their hearing,its working principle is to convert sound signals into electrical impulses to stimulate the cochlear auditory nerve of the patient,so that they can regain their hearing to some degree.At present,implanters can talk normally in a quiet environment and have a high speech recognition rate.However,in scenes with noisy background,such as streets,shopping malls,restaurants and other noisy scenes,the speech recognition rate of cochlear implanters will drop dramatically.In addition,the manufacturing methods of existing implant electrodes for cochlear is complicated with low yield rate and quite few electrodes,making it impossible to implement a more precise coding strategy,resulting in poor hearing experience for implanters.Speech enhancement algorithms can effectively improve the speech recognition rate of implanters in noisy scenes,but traditional speech enhancement algorithms only have a good suppression effect on one or several specific types of noise,which cannot be applied to all scenarios in daily life.Therefore,an algorithm capable of detecting environmental changes and adjusting speech enhancement parameters according to the detection results is needed.Based on the existing theory,this paper proposes a cochlear speech enhancement algorithm based on environmental classification.This algorithm can automatically adjust the noise suppression parameters to adapt to different noise environments.The simulation experiment uses the improved spectral subtraction method as a control.Firstly,this paper tested the performance of the proposed speech enhancement algorithm in three different scenarios,and then the effect is compared with the improved spectral subtraction method in the same scenario.The simulation results show that compared with spectral subtraction,the speech enhancement algorithm based on environmental classification has better performance in saving the voice information while suppressing the noise,especially when the signal-to-noise ratio is-5dB to l0dB,at the same time,the quality of enhanced speech is improved.In terms of electrodes,a large number of studies have pointed out that the current number of electrodes cannot achieve a more precise coding strategy,and new electrode preparation processes need to be proposed to increase the number of electrodes.The article starts with introducing the performance parameters and manufacturing methods of existing electrodes,analyzing the shortcomings of existing electrodes,and proposes a new method for preparing cochlear implant electrodes based on the shortcomings of existing electrodes.The method proposed in this paper first uses the electron beam evaporation technology to produce a flat film,and then rolls the film into a flexible electrode.It is relatively easy to increase the number of electrodes on the basis of points,for example,44,which greatly increases the number of contacts and provides the possibility for patients to listen to music.