Automatic Transcription of Piano Music Based on Convolutional Neural Networks and Convolutional Networks for Biomedical Image Segmentation Model

With the rapid development of intelligent algorithm, its powerful characteristic signal extraction ability has been widely used in the field of music transcription. Aiming at the problems of low recognition accuracy and long transcription time of existing transcription methods, a gated cyclic unit convolutional neural network algorithm was proposed, and a new music transcription model was constructed based on Convolutional Networks for Biomedical Image Segmentation model. The experimental comparison between the proposed algorithm and the traditional algorithm shows that the accuracy of the proposed algorithm is $97.3 \%$, and the loss value is $\mathbf{0. 0 1 9 \%}$, which is significantly better than the comparison algorithm. The results show that the accuracy error of the research model is as low as $0.023 \%$, and the response time is only 22 ms, which is significantly better than the comparison model. The above data fully reflects the powerful music transcription ability of the research model, effectively solves the problems existing in the existing transcription methods, and is conducive to timely and accurate transcription of piano music.