Embedded Systems for Analyzing Digital Art Aesthetics in Piano Performances using Emotional Recognition

Abstract

Aesthetics is an innate ability. It is a meaningful study to make computers perceive "beauty", discover "beauty" and generate "beauty". With the deepening of intelligent optimization algorithm research, artificial intelligence technology "aesthetic" has penetrated into photos, paintings, web pages, ICONS, men and other aspects. However, there are very few studies on the evaluation of piano performance aesthetics. The study of piano performance aesthetics has certain research significance. First of all, limited by personal time and energy, people cannot select high-quality piano repertoire quickly. Secondly, limited by personal aesthetic consciousness and aesthetic ability, people cannot improve the aesthetic quality of piano music just like professional piano players. In the face of such problems, the aesthetic quality evaluation and improvement technology with artificial intelligence as the core provides economically feasible solutions for people to obtain high-quality tracks. Meanwhile, this technology promotes the development of simulated human aesthetic and thinking technology in the field of artificial intelligence. Since the key to aesthetics lies in the perception and classification of piano music score, timbre, audio and emotion, the emotion recognition of piano performance is crucial for the research of artificial intelligence "aesthetics". Piano performance emotion recognition is realized by using the computer to analyze performance characteristics and according to the mapping relationship between performance characteristics and emotion. The study of automatic emotion recognition of piano performance is of great significance to improve the human-computer emotional interaction ability of computer. Based on the above analysis, the main work and innovations of this paper are as follows:This paper first with MIDI music file as a research sample, follow the research method of classical music theory, combined with music psychology, cognitive psychology, music aesthetics and other related research results, the characteristics of the piano performance of a comprehensive and detailed description, and established a set of suitable for computer understanding and expression of the piano performance characteristics system. In the process of feature extraction of piano performance features, high-level features such as rhythm, speed and melody are mathematically defined. In this paper, we realize the computer recognition of the piano playing emotion by using the BP neural network. Finally, the research in this paper can realize the emotional classification of piano performance from the perspective of artificial intelligence, which can use the above research content to quickly and automatically select high-quality piano performance tracks, saving a lot of time for manual screening.