Knowledge and data co-driven intelligent assessment of Chinese zither fingerings

Abstract

The intelligent assessment of musical instrument fingerings can provide learners with timely feedback to greatly improve learning efficiency and lay the foundation for distance teaching. This paper first proposes knowledge and data dual-driven evaluation solution for Chinese zither (Zheng) fingerings, to the best of our knowledge, by integrating the Zheng professional knowledge, fingering video data and intelligent assessment methods. Firstly, we design an experimental paradigm and formulate an assessment scale to ensure the professionalism and intelligence of fingering evaluation. Moreover, due to the lack of musical instrument fingering datasets, a basic fingering dataset with interpretability is established for the zither right hand (CF-Dataset). In order to analyze the fingering movement more, this paper finally proposes a fingering assessment method based on Zheng prior knowledge. The experimental results show that the pipeline designed in this paper is feasible and effective, which makes a pioneering exploration for the combination of musical instrument professional fingering and intelligent video assessment.