Automatic music recognition is an emerging interdisciplinary field with significant applications in music retrieval and automatic music composition. In the context of piano note recognition, the goal is to automatically convert music audio files into digital music files in the MIDI format. A critical component of this process is accurate note endpoint detection, which serves as the foundation for tone recognition. While current research in music recognition primarily focuses on recognizing single notes, there are limitations in terms of recognition accuracy and resistance to noise. In this paper, an intelligent interactive teaching approach is proposed. E-education, as a key technology in internet applications, has become increasingly important in the field of education. Intelligent interactive teaching exhibits several key features, including distributed computing and centralized management. Users access the system through web browsers, allowing the system to provide services to users anywhere with internet access. Furthermore, the system possesses reasoning capabilities, enabling it to analyze user interests more accurately, offer solutions, and address problems in a targeted manner.