Related Experiment Video
Updated: Jan 11, 2026
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
999
A deep learning framework for objective aesthetic evaluation of indoor landscapes using CNN-GNN model
Zixuan Zheng1,2, Ding Yang3, Li Zeng4
1College of Furniture and Art Design, Central South University of Forestry and Technology, Changsha, 410004, China.
Scientific Reports
|November 19, 2025
Summary
This study introduces a deep learning framework using convolutional neural networks (CNNs) and graph neural networks (GNNs) for objective indoor landscape aesthetic evaluation. The novel approach significantly improves accuracy and aesthetic scoring compared to traditional methods.
Area of Science:
- Computer Science
- Architecture
- Environmental Psychology
Background:
- Urbanization increases the importance of indoor landscape design for well-being.
- Subjective human judgment limits traditional aesthetic evaluation methods.
- Need for objective and efficient indoor landscape assessment tools.
Purpose of the Study:
- To develop a deep learning framework for objective indoor landscape aesthetic evaluation.
- To integrate CNNs and GNNs for comprehensive feature extraction.
- To enhance the efficiency and accuracy of aesthetic assessments.
Main Methods:
- Utilized a hybrid deep learning model combining CNNs and GNNs.
- Trained and evaluated the model on benchmark indoor landscape datasets.
- Analyzed both global and local aesthetic features from images.
Main Results:
- Achieved 97.74% accuracy, a 7.54% improvement over conventional methods.
- Demonstrated a 14.21% higher aesthetic score.
- Showcased a 10.6-point improvement in functional evaluation metrics.
Conclusions:
- The CNN-GNN framework offers a robust and objective solution for indoor landscape aesthetics.
- This approach facilitates efficient design optimization and user experience enhancement.
- Highlights the potential of AI in architectural and environmental design evaluation.
