Feature Recognition of Musical Notes Based on Feature Matching Algorithm

Note recognition technology has important application value in the establishment, management and indexing of music database. How to accurately segment notes is a difficult problem in retrieval system. Music is an important mean for human beings to express their feelings and transmit information to each other. It is the embodiment of human wisdom and perceptual thinking. Many thoughts and feelings that cannot be accurately described in words can be expressed through music, so music is an important part of people’s ideological exchange. In this paper, a feature-based matching algorith m is proposed based on music note feature recognition. Firstly, the likelihood value of each group of candidate matching points is estimated by the feature matching model, which is used as the matching probability value between the next group of candidate points and all other candidate points in the feature matching model. The feature matching algorithm has good accuracy and adaptability for random point matching and note feature recognition, and can resist the interference of geometric distortion and singular points. The algorithm in this paper is applied to target location and recognition, and it also achieves good results and has good practical value.