Music sheet score recognition of Chinese Gong-che notation based on Deep Learning

Gong-che notation (GCN) is a major recording method for traditional Chinese music with a long history, first appearing in the late Tang dynasty and the period of the Five Dynasties in China. After an extensive developmental period, it matured during the Ming and Qing dynasties in China and became widely used in traditional Chinese music such as Chinese Kunqu opera, which has a rich heritage. In 2001, it was the first to be declared as a “Masterpiece of the Oral and Intangible Cultural Heritage of Humanity” by UNESCO. A GCN musical score contains meaningful semantic information, including Puzi, lyrics, tune, name, author name, tonality, and notes that are organized in many different styles depending on the particular composition. Collections of the score also contain artifact noise, such as stains and borders. This paper presents a multi-layer integrated classification network to extract semantic information from GCN musical scores by following traditional Chinese music theory combined with an analysis of spatiotemporal information and GCN criterion. The network consists of a multi-layer classifier, which identifies different classification objects by applying various different artificial intelligence methods at each level via clustering analysis, Bayes decision theory, genetic algorithm, deep learning, etc. According to the semantic assessment of the music information of the GCN musical score, this project proposed a similarity measurement method by comparing the manual annotation results and then optimizing the multi-layer integrated classification network. This paper studies the information mining of Chinese traditional musical notation using an interdisciplinary approach and presents a series of methods and schemes to develop and transmit Chinese traditional music in the digital age. It can help protect the proclaimed masterpieces of the oral and intangible heritage of humanity and promote the development of Chinese traditional art and arts technology.