Automated Identification of Tourist Activities in Social Media Photographs: a Comparative Analysis Using Visual-based, Textual-based and Joint-based Methods
Keywords:tourist activity, image identification, Instagram, social media photography, data fusion
The studies pertaining to image identification of tourist photographs are mainly dealing with objects/landscapes, while the activities of tourists interacting with these objects is not well researched. The eligible methods to identify in-depth activities are likewise greatly missing. In this paper, we first explore the feasibility of using different data approaches (visual and textual) to identify tourist activities in social media photos. We further develop a multimodal method combining both text-based and visual-based information. The performances of these methods are compared and validated by manual reviewing. The findings confirm that data fusing methodology is improving identification of micro-level activities.
How to Cite
e-Review of Tourism Research (eRTR) is an international electronic bulletin for tourism research (ISSN:1941-5842). It comprises current tourism research articles, commentaries and reviews by industry professionals. The materials are provided for the personal noncommercial use of registered users of the eRTR, free to individuals and institutions. Copies of articles may be distributed for research or educational purpose, free of charge and without permission. However, commercial use of the eRTR or the articles contained herein is expressly prohibited without the written consent of the publisher.
In consideration for publication of your work, if published on behalf of the eRTR, the author agrees to transfer the work to the eRTR, Department of Recreation, Park and Tourism Sciences, Texas A&M University, USA, including full and exclusive rights to publication in all media now known or later developed, including but not limited to electronic databases.
The authors represents and warrants:
- That the manuscript submitted is his/her own work;
- That the work submitted to the eRTR has not been previously published.