A Social Framework for the Organisation and Automated Annotation of Personal Photo Collections

Abstract

Any large library of information requires efficient ways to organise it and methods, which allow people to access information efficiently. Collections of digital images are no exception. In this proposal we present a framework based on social interaction, to automatically create rich and accurate tags for images using state-of-the-art computer vision techniques, fused with the wide range of context information that is typically available about an image. This framework will enable new and novel ways to organise and present images to users. Due to the unique way that the images are organised many types of queries that would not be possible using standard information retrieval techniques can be processed successfully. The main aim of this work is to utilise the current popularity of online social networks, and fuse online social interaction with computer vision techniques to enable better organisation, retrieval and presentation of images from personal photo collections.