Abstract

The complex aesthetic and structural conventions of Chinese calligraphy pose significant challenges for instructors in developing reliable, objective assessment frameworks and in providing precise, actionable feedback. This subjectivity could obscure students’ learning capability and limit their ability to efficiently identify and address areas in need of writing technique/technical improvement. This study aims to approach this issue by establishing an evaluative body of quantifiable character features as a foundation to implementing an AI-based learning mechanism. Designed to provide specific and objective feedback for side-by-side copying, this mechanism is established by benchmarking expert-written samples through digitized, statistical analysis of pixel optics distributions in individual brushstrokes, converting stylistic techniques into quantitative variables.

With the goal of systematically enhancing traditional learning through unbiased feedback, this study applies a STEAMS (Science, Technology, Artificial Intelligence, Math, and Statistics) research framework to the artistry of Chinese calligraphy. Science is integrated through analysis of evolving theoretical principles of Chinese calligraphy and biophysical principles underlying brush control and stroke formation; Technology and Engineering through application of grayscale and RGB values to benchmark model segment patterns. Math and Statistics quantify the evaluation metrics of the AI model through descriptive statistical analysis and visualizations. Results demonstrate comprehensive feedback through stroke segment connections and generalized scoring criteria. This methodology promises potential future applications in other artistic disciplines.