Self-spectacularization of tourists in visual social media: A computer vision and deep learning approach to socio-cultural body schemas

In the digital media age, travel photography and online sharing have become significant means for tourists to document their travel experiences, display identity, and express emotions, emerging as a prevalent phenomenon of self-spectacularization. These digital images not only reflect tourists' travel experiences but also influence their behavioral expressions and identity formation within a socio-cultural context. This study combines computer vision and behavioral analysis to construct a body schema coding system covering the head, arms, legs, torso, and hands. Based on the theory of symbolic interaction, the study analyzes the social significance of body postures in tourist photographs. Based on a dataset of 555,970 images collected from Weibo and Instagram platforms, the study utilizes Residual Neural Networks and MediaPipe technology for pose recognition and feature extraction. Latent Class Analysis is employed to identify tourists' body expression patterns in various travel contexts. The findings reveal three distinct body schema types: emotional, communicative, and experiential, each corresponding to behavioral tendencies such as emotional expression, social interaction, and immersive experience. By combining symbolic interactionism with behavioral analysis, this study uncovers how tourists use body posture to convey emotions and social signals in social interactions and identity construction. This study not only demonstrates the potential application of machine learning techniques in social behavior studies but also offers a novel methodological perspective for understanding tourism behavior based on visual data.