| Summary: | 碩士 === 國立臺北科技大學 === 電子工程系研究所 === 104 === This paper is based on seniors Wen-Te Kuo paper for optical music recognition system (OMR),we will propose a method to improve it. The experiments show that the system can recognize the sheet music successfully.
The digitized music has many benefits generally, such as Performers can use computer accompaniment to exercise ,build the music library to reduce the memory space, editing or printing.Techniques used in this paper: Run length encoding, vertical and horizontal projection, neural network, all these tips will be description in the paper.
The system can be divided into three stages. First, sheet music recognition. Second, music symbols through the neural network to recognize it. The third phase, the program will convert the results into MIDI files, then play it by using software. After many experiments were done, proving success rate averaged 90%.
|
|---|
