Summarizing Results of Keyword Search on Social Photos using Clustering-based Burst Detection

Social photos are getting attention from researchers to extract real world topics, because users on social media have been posting social photos related to not only personal topics, but also social topics. We are developing curation techniques for social photos. This study proposes a new summarization method for results of keyword search on social photos. Social photos are photos that are posted on social media sites and they usually include posted time and text message as well as photos. It is difficult to know about hot topics in results of keyword search on social photos, because, the huge number of results are returned and they are temporally ordered. The propose method can extract topics as clusters and it also can extract their time changing to identify burstiness of clusters using the clustering-based burst detection.