一种音乐风格的识别与分类方法及系统
By unifying the audio sampling rate, extracting multi-dimensional acoustic features, and dynamically matching classifiers, the accuracy problem caused by sampling rate differences in music style recognition and classification is solved, achieving high-precision and reliable music style classification.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- MINNAN NORMAL UNIV
- Filing Date
- 2026-04-10
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, the difference in sampling rate between the target audio and the reference audio library leads to deviations in acoustic feature extraction, which affects the accuracy of automatic music style recognition and classification.
By judging and adjusting the sampling rate consistency, the target audio and the reference audio are unified to the same sampling rate benchmark. Multi-dimensional musical acoustic feature parameters are extracted, and the most suitable classifier type is matched through intelligent optimization algorithm. Combined with credibility evaluation, music style classification results are generated.
It solves the problem of acoustic feature deviation caused by sampling rate differences, ensures the accuracy and reliability of music style classification, improves feature robustness and classification accuracy, and provides a clear credibility assessment.
Smart Images
Figure CN121996811B_ABST