Handwritten Note Recognition in Digital Music Classroom Teaching Based on Trusted Neural Network

Optical music score recognition mainly studies how to form music score images into computer recognizable semantic symbols. Especially for handwritten music score images, it is particularly difficult to recognize and delete spectral lines due to the complexity and variety of handwritten notes and the addition of various deformations and noises. Therefore, the study of spectral line deletion in handwritten music score images has important theoretical significance and practical value. In this paper, a handwritten music score line deletion algorithm based on multi-scale and multi-directional LBP and XGBoost is proposed. LBP features are designed according to the characteristics of music score images. LBP features can be automatically learned through supervised learning. Spectral line deletion is always performed pixel by pixel, and the test takes a long time. Therefore, the parallel design can be used to improve the efficiency of the operation. The results of this paper have certain reference value for handwritten note recognition in digital music classroom teaching based on trusted neural network.