Neuromorphic auditory event-based anti-reverberation processing method and system for security scene abnormal voiceprint recognition

By introducing a channel-level dynamic refractory period mask window and adaptive leakage adjustment of membrane potential into the neuromorphic auditory system, the problem of performance degradation caused by delayed redundant pulses in high reverberation environments is solved, and abnormal voiceprint recognition with high accuracy and low power consumption is achieved.

CN122157695AInactive Publication Date: 2026-06-05YUESHENGDA (TIANJIN) INTELLIGENT TECHNOLOGY CO LTD
0 Cites 1 Cited by

Patent Information

Application Number
CN202610277256.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-09
Publication Date
2026-06-05
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing neuromorphic auditory technologies suffer from reduced recognition performance and increased false alarms due to delayed redundant pulses generated by multipath echoes in highly reverberant enclosed spaces. Current methods cannot effectively suppress these delayed redundant pulses.

Method used

In a purely asynchronous auditory discrete event sequence, the instantaneous pulse excitation density rate of the frequency channel is calculated, the direct real pulse cluster is identified and the frequency envelope baseline vector is generated. The channel-level dynamic refractory period mask window is set in combination with the physical characteristics of the deployment scenario to intercept delayed redundant pulses, and the pulses are classified by a deep cascaded pulse neural network.

Benefits of technology

It effectively filters out redundant pulses generated by physical echoes, improving the accuracy and robustness of abnormal voiceprint recognition, preventing false alarms, and maintaining the advantage of low power consumption.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an anti-reverberation processing method and system based on a neuromorphic auditory event for security scene abnormal voiceprint recognition. The method comprises the following steps: acquiring an asynchronous auditory event sequence in a high reverberation space; anchoring an initial pulse cluster triggered by a direct sound wave and extracting the initial pulse cluster as a reference feature; on the basis of the reference feature and the estimated reverberation duration, constructing a channel-level dynamic unexpected period mask window; in the action period of the mask window, judging a pulse cluster which arrives newly and is similar to the reference feature as a delay ghost pulse generated by echo, and intercepting and filtering the delay ghost pulse so as to generate a pure pulse stream; and inputting the pure pulse flow into the pulse neural network for identification. According to the method, redundant pulses generated by physical echoes are effectively filtered out from a data source in a native asynchronous pulse domain, high-cost waveform reconstruction is avoided, misjudgment and computing power saturation caused by redundant data of a subsequent pulse neural network can be effectively prevented, the accuracy, robustness and real-time performance of abnormal voiceprint recognition of a system in a high reverberation environment are improved, and the system performance is improved. And meanwhile, the advantage of low power consumption of event-driven sensing is reserved.
Need to check novelty before this filing date? Find Prior Art