Unlock instant, AI-driven research and patent intelligence for your innovation.

Audio Data Clustering Method Based on Spectral Clustering

A technology of audio data and clustering method, which is applied in the direction of electrical digital data processing, digital data information retrieval, special data processing applications, etc., can solve problems such as different labels, and achieve the effect of superior clustering effect

Active Publication Date: 2020-03-31
上海企创信息科技有限公司
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the early music classification methods, record companies often added genre labels artificially for buyers to choose, and sometimes websites that specifically recorded music added labels. Different people often have different feelings about the same piece of music, so it is extremely difficult It is possible that a different tag was added

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
  • Audio Data Clustering Method Based on Spectral Clustering
  • Audio Data Clustering Method Based on Spectral Clustering
  • Audio Data Clustering Method Based on Spectral Clustering

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0027] Such as figure 1 As shown, an audio data clustering method based on spectral clustering, first preprocesses the audio data; calculates the audio period of the audio data, performs frame processing according to the audio period, and extracts audio features; The frequency sequence variance of each frame is the horizontal axis, the logarithmic value of the sequence variance of the power sum of each frame is the vertical axis, and the average value of the power sum is the Z axis, and the audio three-dimensional coordinate system is constructed to obtain the three-dimensional audio vector, and then according to the audio vector The distance is used to calculate the similarity to obtain the similarity matrix S of the audio data; finally, a spectral clustering method is designed to obtain the clustering result of the audio data.

[0028] 1. Audio data preprocessing

[0029] To get an ideal clustering result, the preprocessing method is extremely critical, not only requires a ...

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

The invention discloses an audio data clustering method based on spectral clustering, comprising the following steps: calculating the audio period of the audio data, performing frame division processing according to the audio period, and extracting audio features; The variance is the horizontal axis, the logarithmic value of the sequence variance of the power sum of each frame is the vertical axis, and the average value of the power sum is the Z axis, constructing a three-dimensional audio coordinate system to obtain a three-dimensional audio vector, and then calculate the similarity according to the distance between the audio vectors Degree, get the similarity matrix S of audio data; use the spectral clustering method to cluster the audio data. The invention can provide a practical method for automatic classification of massive music, and can accurately recommend to different users to enhance user experience.

Description

technical field [0001] The invention relates to an audio data clustering method, in particular to an audio data clustering method based on spectral clustering. Background technique [0002] In the past two decades, due to the astonishing development speed of the Internet, massive amounts of information have continued to emerge. How to find useful information from massive amounts of information has become a major problem faced by major network data companies. Traditional statistics and calculations can no longer meet the needs of the public and major companies, and methods derived from data mining, machine learning and other fields have developed rapidly. By setting certain rules and conditions, useful information in massive data can be quickly and effectively found. [0003] NetEase Cloud Music has collected 35 million different songs and music. At the same time, in the United States, about 50 albums are released every week, and each album has about 12 pieces of music on a...

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): G06F16/683G06K9/62
CPCG06F16/683G06F18/2323G06F18/22
Inventor 徐森徐秀芳花小朋徐静徐宁皋军安晶曹瑞
Owner 上海企创信息科技有限公司