After a quite intensive week, WWW has concluded and it’s seems a good time to look back and summarize the main activities and experiences. Though it was my intention to keep a dairy log of the events, the lack of time was much more important than I expected. So, let this late entry replace the originally planned ones 😉
Yesterday it was my turn for presenting our full-paper accepted into this year edition of ACM World Wide Web conference (WWW’09). The expectation for the “Photos and Web 2.0” session was far beyond the calculations of the organizers, and the room allocated for the talks was simply too small for all the people that showed up. Many had to see the first presentation of the session looking through the door while standing in the corridor. It was during the second presentation (my one) that they decided to remove the panels at the rear to merge with an empty room just behind. Though it was absolutely necessary I simply do not understand why they decided to do it right in the middle of my talk. I loss my concentration completely and it was difficult to re-start the talk again.
Despite the difficulties, the presentation went alright and it lead to a positive reaction from the audience which asked a good amount of interesting questions. The slides are available from the www2009 epapers website.
See you in Boston next July! Sigir awaits!
WWW started today with the round of tutorials and workshops that commonly precede such large conferences. I met again with some of the people I’ve been running into in the last few conferences I’ve attended and used the chance to receive an update from them. I also met with my former college Stefan Siersdorfer, co-author of the submission I got accepted into this conference.
My submission to SIGIR 2009 has been accepted as a full paper. The article, entitled “Automatic Video Tagging using Content Redundancy”, proposes an interesting approach to the exploitation of redundant content in folksonomies. We consider the specific case of the leading video sharing website, YouTube. CBCR techniques are used to automatically detect duplication in the video collection, and several metadata propagation methods are proposed to spread community knowledge around the graph of resources.
The abstract of the paper follows below:
“The analysis of the leading social video sharing platform YouTube reveals a high amount of redundancy, in the form of videos with overlapping or duplicated content. In this paper, we show that this redundancy can provide useful information about connections between videos. We reveal these links using robust content-based video analysis techniques and exploit them for generating new tag assignments. To this end, we propose different tag propagation methods for automatically obtaining richer video annotations. Our techniques provide the user with additional information about videos, and lead to enhanced feature representations for applications such as automatic data organization and search. Experiments on video clustering and classification as well as a user evaluation demonstrate the viability of our approach.”
My joint work with Stefan Siersdorfer “Ranking and Classifying Attractiveness of Photos in Folksonomies” has been accepted as a full paper in the next edition of ACM Conference on the World Wide Web (WWW’09).
In this paper, we propose a novel methodology to derive a metric of photo attractiveness in a completely automatic manner taking advantage of user generated data in Flickr (namely metadata and user feedback statistics) as well as image visual features. The paper can be downloaded from the conference site using the following link. Slides for the presentation will be soon available from the same site.
Find below the abstract of the submission:
“Web 2.0 applications like Flickr, YouTube, or Del.icio.us are increasingly popular online communities for creating, editing and sharing content. The growing size of these folksonomies poses new challenges in terms of search and data mining. In this paper we introduce a novel methodology for automatically ranking and classifying photos according to their attractiveness for folksonomy members. To this end, we exploit image features known for having signiﬁcant effects on the visual quality perceived by humans (e.g. sharpness and colorfulness) as well as textual meta data, in what is a multi-modal approach. Using feedback and annotations available in the Web 2.0 photo sharing system Flickr, we assign relevance values to the photos and train classiﬁcation and regression models based on these relevance assignments. With the resulting machine learning models we categorize and rank photos according to their attractiveness. Applications include enhanced ranking functions for search and recommender methods for attractive content. Large scale experiments on a collection of Flickr photos demonstrate the viability of our approach.”