MAVEN at the ICAART 2015

During the 7th International Conference on Agents and Artificial Intelligence (ICAART), Luca Piras (research associate at the Pattern Recognition and Applications Lab of the University of Cagliari) presented MAVEN, showing a demonstration tool developed within the project with the following functionalities: text recognition and detection, logo recognition, object recognition, scene recognition and face recognition and detection.

The conference took place in Lisbon, 10-12 january, 2015.

For more information, please visit:

http://www.icaart.org/EuropeanProjectSpace.aspx?y=2015

A forensics workshop supported by MAVEN

The Workshop on Web Multimedia Verification (#WeMuV2015) will take place in Torino (Italy) the next June 29, co-located with the 2015 IEEE International Conference on Multimedia and Expo (ICME 2015). It is supported by MAVEN and two other EC-funded projects:

More information about the event and instructions for paper submission follow below:

Workshop page: https://sites.google.com/site/wemuv2015/home
Submission deadline: 30 March 2015

Scope & Topics

The wide availability of multimedia capturing equipment (e.g. smartphones, cameras) and the ever increasing use of online social networking and media sharing platforms such as Twitter and Facebook have led to the pervasive use of multimedia content, often user-generated, for reporting on and documenting news stories and events. Establishing the authenticity and veracity of online multimedia content appears to be an increasingly challenging problem, especially in settings where time is a very scarce resource (e.g. breaking news reporting) and the amount of data to be analyzed increasingly large. Failure to verify multimedia content may have severe consequences ranging from personal and brand reputation damage to widespread panic among civilians in the case of natural disasters.

It becomes clear that the availability of methods and tools for automatically assessing the veracity and authenticity of multimedia content published online is of primary importance for ensuring reliable and objective information of the public. Yet, developing such methods in the era of pervasive multimedia capturing and sharing presents a set of unique challenges. Notably, there has been an increasing trend for commoditization and sophistication of media editing software and services, sometimes to the point of being seamlessly intertwined in the media generation process; e.g. several digital filters are at the disposal of Instagram users immediately after the capturing of a photo, leading to the widespread online availability of digitally forged multimedia (be it for malicious or entertainment purposes). What is more, once a multimedia item is published (e.g. in the context of a breaking news story) and attracts attention, it is shared, commented and actively scrutinized by a diverse community of online users, giving rise to new opportunities for gathering insights regarding its veracity. Leveraging such content and its surrounding online context holds the potential for new effective means of verification. What is more, novel computing paradigms putting the human-in-the-loop, e.g. crowdsourcing and collaborative information management platforms, offer new opportunities for leveraging human intelligence at scale for tasks that are hardly solvable by existing algorithmic solutions.
To this end, the proposed Workshop on Web Multimedia Verification aims to highlight the research challenges and new problems arising in the emerging setting of Social and Mobile Multimedia and calls for new approaches, studies and evaluation methodologies and resources that could contribute to more effectively address the research challenges at hand.

In particular, the Workshop calls for submission addressing, but not limited to, the following topics:
– image and video forgery detection
– near- and partial duplicate image detection in image verification problems
– video copy detection in video verification problems
– image and video phylogeny and memes
– audio forensics
– contextual multimedia verification
– fake multimedia propagation and tracking in social media
– multimodal geographic location verification of Web multimedia
– fake multimedia real-world case studies and analyses
– social computing for multimedia verification (e.g. crowdsourcing)
– evaluation methodologies, datasets and benchmarks

Across all topics, particular emphasis should be paid on the real-world aspects of the problem at hand; for instance, for conducting experimental evaluations real cases of content forgeries will be favoured over synthetic ones, datasets from Web and social media sources are of high interest and analysis of applicability to real cases and limitations will be greatly appreciated.

Submission Instructions and Proceedings

Workshop papers should be up to 6 pages long and comply with the guidelines of the ICME 2015 papers: http://www.icme2015.ieee-icme.org/authorguide.php
Submission deadline is 30 March 2015.
The review process will be double-blind. Please prepare your paper in a way that preserves anonymity of the authors.
All ICME workshop papers will be included in separate proceedings in IEEE.

Important dates
30 Mar 2015    Workshop paper submission deadline
30 Apr 2015    Notification of review results
15 May 2015    Camera ready paper submission
29 Jun 2015    Workshop Day

Organizing Committee

Symeon Papadopoulos, CERTH-ITI, Greece
Mauro Barni, University of Siena, Italy
Alessandro Piva, University of Florence, Italy
Alan Hanjalic, TU Delft, The Netherlands