The impact of visual emotional cues in cultural heritage on public sentiment and behavioral intention: an image emotion recognition approach

Abstract This study developed the Heritage Sentiment Index (HSI) using deep learning to detect emotions in cultural heritage images on social media. By analyzing images of historical sites, intangible heritage, and artifacts, this research created a daily sentiment series linked to tourism intentions and user interactions. Results show that HSI accurately predicts public emotional responses and offline participation, especially during sensitive periods such as cultural disputes and disasters. Comparing image-based sentiment (HSI) with comment-based sentiment (CSI) reveals a dual-path process: images prompt immediate reactions, while comments influence delayed responses. When aligned, they amplify emotions; when competing, substitution occurs. This research highlights the emotional impact of heritage imagery in cultural tourism, offering insights for improving digital heritage experiences, multimodal content design, and visitor engagement strategies.