Evolution and Stylistic Characteristics of Ancient Chinese Stone Carving Decoration LSTM-DL Approach with Image Visualization

Yuan TianInternational Journal on Recent and Innovation Trends in Computing and Communication 2023 · DOI

Abstract

In recent years, advancements in data analysis techniques and deep learning algorithms have revolutionized the field of art and cultural studies. Ancient Chinese stone carving decoration holds significant historical and cultural value, reflecting the artistic and stylistic evolution of different periods. This paper explored the Weighted Long Short-Term Memory Deep Learning (WLSTM – DL) evolution and stylistic characteristics of ancient Chinese stone carving decoration through the application of image visualization techniques combined with a Long Short-Term Memory (LSTM) time-series deep learning architecture. The WLSTM-DL model uses the optimized feature selection with the grasshopper optimization for the feature extraction and selection. By analyzing a comprehensive dataset of stone carving images from different periods, the WLSTM-DL model captures the temporal relationships and patterns in the evolution of stone carving decoration. The model utilizes LSTM, a specialized deep-learning architecture for time-series data, to uncover stylistic characteristics and identify significant changes over time. The findings of this study provide valuable insights into the evolution and stylistic development of ancient Chinese stone carving decoration.

Introduction

Ancient Chinese stone carvings serve as rich repositories of historical, religious, and social narratives. For centuries, the study of these artifacts has relied almost exclusively on the manual expertise of art historians. While invaluable, this approach is subjective and difficult to scale across the vast corpus of surviving artifacts. The primary challenge lies in quantifying "stylistic drift"—the subtle transitions in motifs and techniques that occur across dynasties.

An example of an ancient Chinese stone carving, illustrating the intricate geometric patterns and symbolic motifs that the WLSTM-DL model aims to classify.

This paper addresses the need for a computational framework that can objectively extract discriminative visual features and model their temporal evolution. By framing art history as a sequence of stylistic developments, the researchers propose a novel deep learning framework: WLSTM-DL.

Methodology

The proposed pipeline integrates metaheuristic optimization with sequence modeling to decode artistic evolution.

Feature Extraction and Optimization

To handle the high dimensionality of image data and the visual noise common in weathered stone, the framework employs the Grasshopper Optimization Algorithm (GOA). Inspired by the foraging behavior of grasshopper swarms, this algorithm identifies the most "informative" features from the extracted image descriptors. This step is crucial for preventing overfitting and ensuring the model focuses on the most stylistically relevant visual markers.

Weighted LSTM-DL Architecture

The core innovation is the application of Weighted Long Short-Term Memory (WLSTM) networks to artistic data. While LSTMs are typically used for text or speech, they are utilized here to capture "long-term dependencies" in artistic trends. By processing features chronologically, the model "learns" how specific stylistic elements—such as line work or motif complexity—evolve from the Tang Dynasty through to the Modern Era.

The architecture of an LSTM cell, featuring gates that control the flow of information to capture long-term temporal dependencies in stylistic evolution.

Key Results and Discussion

The WLSTM-DL model was evaluated against a chronologically organized dataset and compared with standard machine learning benchmarks like Support Vector Machines (SVM) and Decision Trees (DT).

  • High Accuracy: The proposed model consistently achieved an accuracy range of 98% to 99%, significantly outperforming traditional methods.
  • Robustness: The model maintained high Precision and Recall (0.96–0.99), indicating its ability to distinguish between the intricate and often subtle differences between late-dynasty styles (e.g., Ming vs. Qing).
  • Interpretability: The integration of image visualization techniques allows the model's internal predictions to be mapped back to visual patterns, providing a quantitative backbone to qualitative historical theories.

Conclusion

The application of the WLSTM-DL model showcases the potential of deep learning in uncovering hidden narratives within ancient artworks. By treating stylistic evolution as a time-series problem, the framework provides a robust, objective method for understanding the intricate details of cultural heritage. Future work will focus on expanding the dataset to include a wider variety of global cultural artifacts to test the generalizability of this temporal stylistic logic.

Summary updated

Abstract

In recent years, advancements in data analysis techniques and deep learning algorithms have revolutionized the field of art and cultural studies. Ancient Chinese stone carving decoration holds significant historical and cultural value, reflecting the artistic and stylistic evolution of different periods. This paper explored the Weighted Long Short-Term Memory Deep Learning (WLSTM – DL) evolution and stylistic characteristics of ancient Chinese stone carving decoration through the application of image visualization techniques combined with a Long Short-Term Memory (LSTM) time-series deep learning architecture. The WLSTM-DL model uses the optimized feature selection with the grasshopper optimization for the feature extraction and selection. By analyzing a comprehensive dataset of stone carving images from different periods, the WLSTM-DL model captures the temporal relationships and patterns in the evolution of stone carving decoration. The model utilizes LSTM, a specialized deep-learning architecture for time-series data, to uncover stylistic characteristics and identify significant changes over time. The findings of this study provide valuable insights into the evolution and stylistic development of ancient Chinese stone carving decoration. The application of image visualization techniques and the WLSTM-DL model showcase the potential of data analysis and deep learning in uncovering hidden narratives and understanding the intricate details of ancient artworks.
BibTeX
@article{tian2023evolution,
  author = {Yuan Tian},
  title = {{Evolution and Stylistic Characteristics of Ancient Chinese Stone Carving Decoration LSTM-DL Approach with Image Visualization}},
  journal = {International Journal on Recent and Innovation Trends in Computing and Communication},
  year = {2023},
  doi = {10.17762/ijritcc.v11i6.7723},
}
Figure