Research on the Clustering of Emotional Elements in Art and Design Based on Visual Language Communication

Abstract

Since many different emotional expressions can be found in artistic and design compositions, it can be challenging to effectively extract and analyze emotional information from such a wide range of artworks. This study uses deep learning approaches to extract and cluster emotional aspects from art and design works to address this problem. The suggested method uses DBSCAN (Density-Based Spatial Clustering of Applications with Noise) for clustering and combines the VGG-16 and Bi-LSTM models for feature extraction from images. The suggested approach works better than existing models in extracting emotional information pieces, according to experimental results. With a Macro-F1 assessment score of 0.9241, the suggested technique can efficiently examine emotional inclinations in artistic and design works in practical applications. In conclusion, this study discusses the potential applications of the suggested emotion element extraction and clustering method in the field of emotional analysis in art and design, offering fresh approaches to issues in this area.