Saturday 08 March 2025
For centuries, art historians and scholars have struggled to fully understand the complex meanings behind ancient Chinese paintings. These works of art are more than just visually stunning; they hold secrets about the cultural, philosophical, and historical contexts in which they were created. However, until now, there has been no unified framework for describing these paintings in a way that does justice to their richness.
A team of researchers from Renmin University of China has developed a revolutionary semantic descriptive model (SDM) that is designed to change the game. By integrating iconological theory with advanced computational methods, the SDM can automatically extract subject terms and construct a user-friendly framework for describing ancient Chinese paintings.
The researchers started by collecting over 1600 ancient Chinese painting collections from the Beijing Palace Museum, along with related documents from the CNKI database. They then used a two-stage automatic approach to extract subject terms from the abstracts of these documents. The first stage involved coarse-grained term recognition, using word segmentation and lexicon analysis to identify relevant nouns and adjectives. In the second stage, term ranking was conducted using an advanced deep learning-based model called EmbedRank.
The extracted subject terms were then clustered based on their representation similarity using the K-means algorithm, resulting in multiple subject clusters with similar semantic terms. These clusters were mapped to a three-layer structure of the SDM, which includes pre-iconographical elements (basic information describing natural subjects), iconographical elements (conventional interpretations of objects), and specific subject categories/subjects.
To validate the effectiveness of the SDM, the researchers conducted a user evaluation test with 8 professional scholars. The participants were asked to assess the usefulness of two systems: one that used the SDM and another that relied solely on basic information provided by the Beijing Palace Museum. The results showed that the SDM significantly outperformed the baseline system in various aspects of usefulness, including diversity, comprehension, effectiveness, and satisfaction.
The implications of this research are far-reaching. For art historians and scholars, the SDM provides a powerful tool for analyzing and understanding ancient Chinese paintings in a more nuanced and comprehensive way. The model can also be used to support further art-related knowledge organization and cultural exploration of these paintings. Furthermore, the advanced computational methods developed in this study have broader applications in various fields, including natural language processing and information retrieval.
In essence, the SDM represents a major step forward in our ability to understand and describe ancient Chinese paintings.
Cite this article: “Unlocking the Secrets of Ancient Chinese Paintings with AI-Powered Semantic Descriptive Model”, The Science Archive, 2025.
Ancient Chinese Paintings, Semantic Descriptive Model, Art History, Computational Methods, Iconological Theory, Natural Language Processing, Information Retrieval, Beijing Palace Museum, Cnki Database, Deep Learning-Based Model.
