Article: Design and practice of artificial intelligence-driven piano improvisation accompaniment teaching system introduction Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.52 pp.56 - 74 Abstract: In this study, we provide a system that shows students how to play piano with improvisation and accompaniment using cloud computing, deep learning, and CNN. Automatic evaluation of performance aspects, such as pitch, timbre, articulation, rhythm, and dynamics, is one way the suggested approach enhances piano lessons. Applying a hybrid approach that combines a matched filter with a rapid guided filter optimises preprocessing for music feature extraction. To further improve the accuracy of piano performance analysis, attention-induced multi-head CNNs and perceptual evaluation datasets are employed. In adaptive and remote learning settings, the technique shows better dependability and scalability. The model successfully integrates visual and aural methods of teaching piano, supports multilevel perceptual feature analysis, by providing a novel framework that enhances learning outcomes, enables tailored instruction, and adapts to the diverse needs of learners, this research contributes to the expanding field of intelligent music education. Inderscience Publishers - linking academia, business and industry through research

Title: Design and practice of artificial intelligence-driven piano improvisation accompaniment teaching system introduction

Authors: Xiang Wei

Addresses: College of Architecture and Arts, Taiyuan University of Technology, Taiyuan, Shanxi, 030024, China

Abstract: In this study, we provide a system that shows students how to play piano with improvisation and accompaniment using cloud computing, deep learning, and CNN. Automatic evaluation of performance aspects, such as pitch, timbre, articulation, rhythm, and dynamics, is one way the suggested approach enhances piano lessons. Applying a hybrid approach that combines a matched filter with a rapid guided filter optimises preprocessing for music feature extraction. To further improve the accuracy of piano performance analysis, attention-induced multi-head CNNs and perceptual evaluation datasets are employed. In adaptive and remote learning settings, the technique shows better dependability and scalability. The model successfully integrates visual and aural methods of teaching piano, supports multilevel perceptual feature analysis, by providing a novel framework that enhances learning outcomes, enables tailored instruction, and adapts to the diverse needs of learners, this research contributes to the expanding field of intelligent music education.

Keywords: artificial intelligence; AI; piano teaching; improvisation accompaniment; convolutional neural network; CNN; deep learning; DL; cloud computing; perceptual features; music education technology.

DOI: 10.1504/IJICT.2025.151170

International Journal of Information and Communication Technology, 2025 Vol.26 No.52, pp.56 - 74

Received: 28 Aug 2025
Accepted: 16 Sep 2025
Published online: 15 Jan 2026 *

Title: Design and practice of artificial intelligence-driven piano improvisation accompaniment teaching system introduction

Keep up-to-date