Aesthetic Evaluation System for Calligraphy Characters using Convolutional Neural Networks

Chinese calligraphy possesses profound historical significance as a form of artistic expression, where the integration of technique and aesthetics plays a crucial role in its lasting impact. The amalgamation of calligraphic technique and aesthetics is essential, as it reflects cultural heritage and artistic excellence. Understanding and evaluating these elements is vital for both practitioners and enthusiasts, as it provides insights into the nuances of calligraphic artistry and fosters appreciation for this traditional art form. This paper introduces a convolutional neural network-based system designed to evaluate the aesthetic quality of Chinese calligraphy. By utilizing the Canny algorithm and Harris corner detection, our system identifies three fundamental aesthetic characteristics: strokes, skeleton, and brushwork, which are essential for calligraphy experts. We classify generated regular and official scripts into professional and amateur fonts created by four distinct users. Through objective analysis, we investigate the interplay between brushwork and ink lines in calligraphic artistry, while examining the aesthetic attributes across various font styles. Experimental results indicate an average accuracy of approximately 91.9%, highlighting the effectiveness of our system in discerning and recognizing the nuances of calligraphy, thus providing valuable insights for users.