Research on the Correlation Between Tourists’ Facial Emotion Perception and Street Spatial Elements Through Deep Learning: A Case Study of Traditional Tourism-Oriented Villages in China

Abstract

The tourism-driven revitalisation of traditional villages has become a significant pathway to achieving historical and cultural heritage preservation and promoting rural revitalisation in the new era. Among these efforts, constructing the regional characteristics of the built environment in villages is a critical aspect of revitalisation. Tourists’ emotional perceptions during their spatial experiences in tourism villages serve as an important evaluation metric for the localization characteristics of the space. This study investigates the correlation mechanisms between 13 spatial elements—such as spatial organization, morphology, and interface—and tourists’ emotional types and intensity preferences based on the distribution characteristics of tourists’ emotional perceptions in the street spaces of tourism villages. Facial expression images of tourists in different village streets were collected via smartphone photography, and a deep learning model, YOLOv5, was trained using a manually labeled dataset to recognize and quantify four types of emotions: smile, surprise, calm, and negative emotions. To enhance recognition accuracy, the model’s performance was improved using techniques such as RGB value adjustment, flipping, and augmentation. Subsequently, a multiple regression analysis was conducted to identify the street space elements influencing tourists’ emotional perceptions. The results indicate that the proposed model optimization methods effectively identified tourists’ emotions under various scenarios with high accuracy. Regression analysis revealed that spatial segmentation factors of streets had a regression coefficient of 0.184, making it a key factor affecting tourists’ perceptions. The findings identified five typical street space elements that positively influence tourists’ emotions. Based on these results, recommendations for optimizing street space to enhance tourists’ perceptions were proposed, offering support for improving street space tourism experiences and promoting the sustainable development of traditional villages.