System for building music classification model, system for recommending music and corresponding method
A classification model and music technology, applied in the computer field, can solve problems such as the unscientific music recommendation effect of the music classification model, and achieve the best experience effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0034] Embodiment 1. This embodiment provides a system for establishing a music classification model, see figure 1 As shown, it includes: a first feature extraction unit 101 , a first feature splicing unit 102 , and a model training unit 103 .
[0035] Wherein, the first feature extraction unit 101 is configured to extract acoustic feature vectors of different dimensions for each piece of music in the training data.
[0036] In the training phase, the corresponding music will be selected in advance according to different music styles as the training data. In the training process, the acoustic features of different dimensions extracted by the present invention for training data may include but not limited to acoustic feature vectors of different dimensions formed by any combination of the speed, strength, timbre, and melody of the song. In the embodiment of the present invention In the description of , the simultaneous use of these four acoustic features is taken as an example...
Embodiment 2
[0040] Embodiment 2. This embodiment provides a system for establishing a music classification model, see figure 2 As shown, it includes: a first feature extraction unit 201 , a first feature splicing unit 202 , a first feature dimensionality reduction unit 203 , a model training unit 204 , a model library unit 205 , a user feedback unit 206 and an adaptive adjustment unit 207 .
[0041] Wherein, the description about the first feature extraction unit 201 and the first feature splicing unit 202 is consistent with the first embodiment, and will not be repeated here.
[0042] The first feature dimensionality reduction unit 203 is configured to provide the supervectors of each music to the model training unit 204 after removing the correlation information of each dimensional acoustic feature vector in the supervector obtained by the first feature splicing unit 202 . Specifically, due to the high dimensionality of the supervector (such as: four dimensions), and there will be corr...
Embodiment 3
[0054] Embodiment 3. This embodiment provides a method for establishing a music classification model, see image 3 shown, including the following steps:
[0055] S301. Extracting acoustic feature vectors of different dimensions for each piece of music in the training data.
[0056] In the training phase, the corresponding music will be selected in advance according to different music styles as the training data. Specifically, in the training process, the acoustic features of different dimensions extracted by the present invention for the training data may include but not limited to acoustic feature vectors of different dimensions formed by any combination of the speed, strength, timbre, and melody of the song. In the description of the embodiments of the invention, the simultaneous use of these four acoustic features is taken as an example for description. These four acoustic features describe the style of a song from different perspectives, and songs with similar styles mus...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com