Abstract

Widespread access to photo-taking devices and high speed Internet has combined with rampant social networking to produce an explosion in picture sharing on web platforms. In this environment, new challenges in image acquisition, processing, and sharing have emerged, creating exciting opportunities for research in computer vision and multimedia data mining. I explore several of these interesting problems in my work reconstructing collective storylines from large-scale, unordered, online image collections. In this talk, I begin by introducing the two fundamental challenges faced in this research: modelling the temporal trends of whole image collections and detecting repetitive content among images. I present proposed solutions to each problem, along with interesting applications enabled by the results. First, I discuss our approach for jointly aligning and segmenting a large number of photo streams from multiple users, as a first technical step to uncovering the common storylines of outdoor recreational activities such as scuba diving and horse riding. Second, I present our results for visualizing pictorial stories associated with the competing brands such as sports brands Nike and Adidas. I conclude my presentation with a discussion of future projects to further extend my research.