An Acoustic Event Recognition Method Based on Subspace Representation Learning

A recognition method and subspace technology, applied in character and pattern recognition, computer components, speech analysis, etc., can solve the problem that semantic features cannot take into account the essential content and timing structure of the original signal

Active Publication Date: 2021-04-02
HARBIN INST OF TECH
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to propose an acoustic event recognition method based on subspace representation learning in view of the problem that the semantic feature extraction of the prior art cannot take into account both the essential content and the temporal structure of the original signal in the AER task

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • An Acoustic Event Recognition Method Based on Subspace Representation Learning
  • An Acoustic Event Recognition Method Based on Subspace Representation Learning
  • An Acoustic Event Recognition Method Based on Subspace Representation Learning

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0026] Specific Embodiment 1: The following describes this embodiment in detail. In this embodiment, an acoustic event recognition method based on subspace representation learning includes the following steps:

[0027] Step 1. Signal preprocessing: first, segment the original acoustic event signal into a training set and a test set, then convert the segmented acoustic event signal into a single-channel signal, and finally sample the above-mentioned single-channel signal;

[0028] The division ratio is that the training set accounts for 75% of the total number of acoustic event signals, and the test set accounts for 25% of the total number of acoustic event signals; then the acoustic event signal is converted into a single-channel signal; finally, the acoustic event signal is sampled to generate a sampling frequency It is a single-channel signal at 16000Hz.

[0029] Step 2. Frame-level feature extraction: First, the sampled single-channel signal is divided into multiple audio f...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

An acoustic event recognition method based on subspace representation learning, which relates to the technical field of sound signal processing. In order to solve the problem that the semantic feature extraction in the prior art cannot take into account the essential content and temporal structure of the original signal in the AER task, it includes: Step 1 , signal preprocessing, step 2, frame-level feature extraction, step 3, sub-acoustic event feature extraction, step 4, timing extension of sub-acoustic event features, step 5, overall semantic feature extraction between sub-acoustic event features, step 6, For the recognition of acoustic events, when the present invention extracts semantic features, it can take into account the overall content information and global timing structure of the original signal.

Description

technical field [0001] The invention relates to the technical field of sound signal processing, in particular to an acoustic event recognition method based on subspace representation learning. Background technique [0002] Acoustic events are an important medium for humans to perceive the surrounding environment. Acoustic event recognition (Audio Event Recognition, AER) is the process of using computer to judge the category of acoustic events by simulating the human auditory mechanism. Due to the existence of many potential applications, such as environmental monitoring, noise control, etc., AER has been paid more and more attention by researchers in recent years. [0003] In order to enable computers to have the ability to recognize and understand acoustic events similar to the human ear, extracting semantic features from acoustic event signals is an important and challenging part of the AER task. In general, the semantic feature is essentially the compression and abstrac...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G10L25/51G10L25/45G10L25/24G06K9/62
CPCG10L25/51G10L25/24G10L25/45G06F18/23213G06F18/2411G06F18/214
Inventor 韩纪庆史秋莹罗辉郑铁然郑贵滨
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap