Photos are a special way to tell stories of our best memories and moments. The representation of those photos in appealing physical photo books is highly appreciated by many people. Today, many photos are shared via social networking sites, where people upload their photos and share their stories with their friends. The members of social networks comment on each other's photos, add tags or descriptions and upload new photos of the same events to their albums. While the media of different personal events are available on the social network, there is no easy way to collect and bundle them into a story and print this story as a photo book. We propose an approach to automatically detect media elements that match a query (where, when, what, who) in the user's social network and intelligently arrange and compose them into a printable photo book. We combine content analysis of text and images to automatically and semi-automatically select photos of a specific story. We calculate the probabilities of each two photos to belong to the same event using an Expectation-Maximization algorithm that we propose in order to be able to retrieve them easily when receiving the user queries, and we address the differences between our model and other models that use similar proposed algorithms. People's tags and the interaction between the users and the photos as well as other semantic information are exploited to select important photos that are suitable to create the photo book. The selected photos and derived semantics are then employed to automatically create an appealing layout for the photo book.

Photos are often a means to remember personal events, and the creation of photo albums is the attempt to preserve our memories in a nice book. For a long time people have been creating such photo albums on the basis of prints from analog photos arranged in an album book with scissors and glue and annotated with comments and captions-a tedious task which in these days is getting support by authoring tools and digitally mastered photo books. Relying on the content of others such as printed travel guides, news papers, leaflets, but also friends and family the personal content often has been enriched, enhanced, and completed. This is the starting point of our work: with digital photography and the increasing amount of content-based and contextual metadata of personal photos we can now use this metadata to actually support the targeted and semi-automatic inclusion of interesting, related information from content of others, e. g., from Web 2.0 communities, and offer and add it at the right spot in the personal album. In this paper, we show how photo album creation can benefit from leveraging information learned from many users in regard of the album's content, structure, and semantics.

No abstract

No abstract

Facebook witnesses an explosion of the number of shared photos: With 100 million photo uploads a day it creates as much as a whole Flickr each two months in terms of volume. Facebook has also one of the healthiest platforms to support third party applications, many of which deal with photos and related events. While it is essential for many Facebook applications, until now there is no easy way to detect and link photos that are related to the same events, which are usually distributed between friends and albums. In this work, we introduce an approach that exploits Facebook features to link photos related to the same event. In the current situation where the EXIF header of photos is missing in Facebook, we extract visual-based, tagged areas-based, friendship-based and structure-based features. We evaluate each of these features and use the results in our approach. We introduce and evaluate a semisupervised probabilistic approach that takes into account the evaluation of these features. In this approach we create a lookup table of the initialization values of our model variables and make it available for other Facebook applications or researchers to use. The evaluation of our approach showed promising results and it outperformed the other the baseline method of using the unsupervised EM algorithm in estimating the parameters of a Gaussian mixture model. We also give two examples of the applicability of this approach to help Facebook applications in better serving the user.