This document summarizes a research paper that extends the PhotoTOC algorithm to cluster photos of social events using richer contextual information. The approach first sorts photos temporally and splits them into mini-clusters, then sequentially merges mini-clusters based on weighted similarities across time, location, tags, and user information. Experiments on the ReSEED dataset show the approach can group related photos while handling diversity in users, locations, and qualities, though it may still incorrectly split or merge some clusters. Future work is needed to optimize parameters, improve the fusion approach, and make criteria event-dependent.