Abstract: With the advent of the Internet + era, traditional piano teaching is facing unprecedented opportunities and challenges. In order to realize the intellectualization of piano network teaching, this paper constructs a piano note recognition algorithm. According to MIDI files and corresponding piano audio files, the algorithm builds training and test waveform data to calculate the corresponding annotation files, and recompiles the multi-note model to realize the multi-note recognition system. The multi tone HMM acoustic model and multi note model are established by using Internet technology. The matching degree between the test audio and the multi note models is calculated respectively, and then an Internet + piano intelligent network teaching system model is constructed. The test results show that when the number of states of the multi-note recognition system is 7, the correct recognition rate is the highest. Through this study, the piano intelligent network teaching system model has produced new inspiration, which can be more accurate when the multi-note recognition state is 7.
Keywords: Internet+, network teaching, intelligent system, piano
