Auditory scene recognition method for artificial cochlea

Abstract

The invention discloses an auditory scene recognition method for an artificial cochlea. The method comprises the steps of (A) establishing a standard scene training UBM, (B) performing framing and windowing on a sound signal, (C) identifying a pre-processed sound signal by frame, (D) performing feature vector extraction on a scene noise signal which is subjected to VAV processing, and (E) processing the signal which is subjected to feature extraction in a GMM-UBM system to obtain a likelihood score and finally identify the type of a scene. According to the method, by establishing a series of models, different auditory scenes can be identified, instructions are provided for signal processing modules of subsequent speech enhancement, speech strategy and the like of a speech processor, therefore, the signal process of the speech processor is more matched with an auditory scene, the clarity and intelligibility of a speech signal of a patient in a noisy environment are improved, at the sametime, the listening effect under a music scene can be improved, and the quality of life of an artificial cochlea implant patient is further improved.

Description

technical field [0001] The invention relates to an auditory scene recognition method, in particular to a cochlear implant auditory scene recognition method. Background technique [0002] Cochlear implants are currently the only effective method and device that can restore hearing to patients with bilateral severe or extremely severe sensorineural deafness. The existing cochlear implant operation process is as follows: the sound is first collected by the microphone and converted into an electrical signal, after special digital processing, and then encoded according to a certain strategy, it is transmitted to the body through the transmitting coil behind the ear, and the receiving coil of the implanted body After the signal is sensed, it is decoded by the decoding chip, so that the stimulating electrode of the implant generates current, thereby stimulating the auditory nerve to produce hearing. Due to the limitation of the use environment, the sound must be mixed with environ...

AI Technical Summary

Problems solved by technology

Due to the limitation of the use environment, the sound must be mixed with environmental noise, and the sound signal needs to be optimized by a certain algorithm. However, in view of the diversification of the use environment, if only a single algorithm is used for optimization, the signal after algorithm optimization may sometimes be different from the actual signal. The situation is deviated, and the best auditory effect cannot be achieved. Therefore, an auditory scene recognition method is needed, so that different scenes use different optimization algorithms to achieve the best auditory effect.

Images