News:


ITI/CERTH establishes cooperation with Hellenic Police/Forensic Sciences Subdivision of Northern Greece

The growth of electronic crime has called for the development of advanced analysis tools that could facilitate the tasks of digital forensic examiners. In this context, ITI/CERTH has established cooperation with Hellenic Police/Forensic Sciences Subdivision of Northern Greece, in order to integrate innovative multimedia technologies into dedicated forensic tools that could support the investigations of electronic crime.

During this cooperation, ITI/CERTH with the guidance of Hellenic Police has developed an interactive multimedia content-based search engine that is capable of retrieving multimedia content and automatically recognizing pornographic material. This search engine is evaluated and tested by the Hellenic Police, in order to be used for facilitating the investigations on child pornography detection.

Improve My City Mobile available on Github under A-GPL

ImproveMyCityMobile is an open source smartphone application for Android devices that allows citizens to report all types of issues related to their neighborhood. It is now available on Github

Multimedia Group @ ICIP 2013

Three papers from the Multimedia Group of the Information Technologies Institute will be presented in the 2013 IEEE International Conference on Image Processing (ICIP 2013), which will be held in Melbourne, Australia on September 15-18, 2013. ICIP, sponsored by the IEEE Signal Processing Society, is among the top international conferences in theoretical, experimental, and applied image and video processing. The papers are:

  • Video Event Recounting Using Mixture Subclass Discriminant Analysis, by Nikolaos Gkalelis, Vasileios Mezaris, Ioannis Kompatsiaris and Tania Stathaki.
  • Enhancing Video Concept Detection with the Use of Tomographs, by Panagiotis Sidiropoulos, Vasileios Mezaris and Ioannis Kompatsiaris.
  • Discovery of Environmental Resources Based on Heatmap Recognition, by Anastasia Moumtzidou, Stefanos Vrochidis, Elisavet Chatzilari and Ioannis Kompatsiaris.

Sense the City

Sense the City is an application that allows citizens to get real time information regarding the environmental conditions in their city. This information is retrieved from several sensors that "sense" the city providing environmental measurements, such as temperature and concentration of pollutants (e.g. CO, CO2, NO2) that affect the air quality.

Sense the City is a web based application that has been developed within the European Project PEOPLE, in close cooperation with the Municipality of Thermi and user groups. It is therefore a truly user-centered digital application, adapted to the needs of Municipalities and local communities. The application is available in Greek and English, and provides air quality and temperature information for the area of Thermi in Thessaloniki. The Information Technologies Institute - Multimedia Group developed this application in cooperation with Draxis company as a subconstractor of the Municipality of Thermi.

Visit this link to sense the city!

The application is a Joomla component, which is open source and available here.

Enviromental Data Storage and Retrieval Service

The Environmental Data Storage and Retrieval Service is a software, which supports storage, spatiotemporal indexing and retrieval of environmental information (e.g. air quality and weather forecasts, observations, historical data) as this is provided by nodes (e.g. sensors, stations, websites) that contain environmental measurements (e.g. weather conditions, air quality, pollen).

This implementation follows the Sensor Observation Service (SOS) 1.0 standard and builts upon 52°North SOS. This service has been implemented in the context of the  European project PESCaDO, which deals with personalized orchestration and delivery of environmental information.

Heatmap Datasets

Heatmaps are graphical representations of matrix data with colors and comprise the most common presentation format of air quality and pollen concentration forecasts. In the context of evaluating the discovery of environmental portals, which include air quaity and pollen forecast data, based on heatmap recognition, CERTH-ITI has created two annotated sets of heatmaps.

Best Paper Honorable Mention at MMM2013, Huangshan, China

The paper "Semi-supervised concept detection by learning the structure of similarity graphs" by Symeon Papadopoulos, Christos Sagonas, Yiannis Kompatsiaris and Athena Vakali won a Best Paper Honorable Mention in the International Conference in Multimedia Modeling that took place in Huangshan, China between 7 and 9 January 2013. The paper was selected out of 96 papers, 43 oral presentations and 53 poster papers that were accepted by the conference (27% acceptance rate for oral, 44% overall acceptance rate). You can find an online presentation of the paper below.

The work for this paper was funded by the SocialSensor Integrated Project.

Improve My City Mobile - Android version

The mobile application enables citizens to report local problems such as potholes, illegal trash dumping, faulty street lights, broken tiles on sidewalks, and illegal advertising boards. The submitted issues are displayed on the city's map. Users may add photos and comments. Moreover, they can suggest solutions for improving the environment of their neighbourhood.

The web based application has been developed within the European Project PEOPLE, in close cooperation with the Municipality of Thermi and user groups. It is therefore a truly user-centered digital application, adapted to the needs Municipalities and local communities. The Android mobile version is developed by the Information Technologies Institute - Multimedia Group.

Visit IMC mobile page for more information

Audio and Video recordings in the Greek Association of Alzheimer's Disease and Related Disorders

Within the framework of Dem@Care and in co-operation with IBM, CERTH and the Greek Association of Alzheimer's Disease and Related Disorders, audio and video recordings have been performed in Thessaloniki with participants from the Greek Association of Alzheimer's Disease and Related Disorders.

The participants were elderly people at the early stage of dementia, with mild cognitive impairment (MCI) and healthy people with similar age range and demographic attributes.

Find out more...

CERTH participation in the Social Event Detection task @ MediaEval 2012

 

CERTH-ITI successfully participated in the Social Event Detection task of the MediaEval 2012 workshop. MediaEval, a workshop focused on focused evaluation tasks with respect to multimedia analysis, took place in Pisa, Italy (4-5 October 2012). CERTH took part in the Social Event Detection (SED) task that involved three challenges: (a) detection of technical events in Germany, (b) soccer events in two major European cities, (c) detection of Indignados events in Madrid. All challenges were defined on a large set of over 160,000 images from Flickr. A more detailed definition of the task is available here.

In the all challenges, CERTH achieved median performance using a versatile and flexible approach developed within the SocialSensor project. The approach employed by CERTH is described here, while the presentation given at the workshop is embedded below. We have also made available the implementation of the approach in the form of a Java library. 

 

Social Event Detection dataset made available for researchers

 

CERTH has successfully contributed to the organization of this year's Social Event Detection task (SED 2012) that took place in Pisa 4-5 October, in the context of MediaEval 2012. The SED 2012 dataset has been made publicly available for download and use by the research community. You can freely download the dataset from http://mklab.iti.gr/project/sed2012 . The downloadable archives include the following:

  • The three 2012 SED Challenges definitions,
  • The XML metadata for the images in the test dataset,
  • The actual image files of the test dataset,
  • Ground truth results for the defined challenges/dataset,
  • Our evaluation script.

Please feel free to use this dataset for research purposes, and also to disseminate the above information to anyone else who may be interested.

Succesful participation of CERTH-ITI in interactive known-item video search of TRECVID 2012

CERTH-ITI participated to the Known-Item Search (KIS) Task of TRECVID 2012 with the interactive video retrieval engine VERGE. VERGE results achieved the second place, among the submissions made by all institutions participating to TRECVID 2012 KIS task (19 runs from 5 different institutions) with interactive systems. The latest version of VERGE supports advanced retrieval functionalities including visual similarity search, visual concept-based retrieval, video similarity based on PLSA and full text search, as well as two different representations (i.e. shot and video based). More information about VERGE is available here. Feel free to try on line the latest version of VERGE and search in the TRECVID 2012 KIS test data collection.

Best moderated poster award at ESC 2012

The paper entitled "Plaques with increased lipid core and thin fibrous cap occur in coronary artery regions with low endothelial shear stress: A 3D Optical coherence tomography pilot study in human", which was presented in the European Society of Cardiology Congress 2012 (ESC 2012), received the "Best Moderated Poster" award. Read more here.

CERTH participated in the ImageCLEF 2012 Flickr photo annotation task

 

CERTH participated in ImageCLEF Photo Annotation and Retrieval 2012, and more specifically in the first subtask, namely Visual and Concept detection, annotation using Flickr photos. The objective of taking part in the competition was to evaluate two multimedia indexing approaches developed within SocialSensor and to compare the performance of methods using different sets of image features.
 
The first of the tested approaches constructs a similarity graph that includes both train and test images and trains concepts detectors using the graph Laplacian Eigenmaps (LE) as features. The second approach utilizes the concept of a “same class” model which takes as input the set of distances between the image to be annotated and a reference item representing a target concept, and predicts if the image belongs to the target concept.
 
Five runs were submitted. Run 1 ranked 5th out of 17 text-based methods (MiAP) and 7th (GMiAP, F-ex). The performance of the visual-based run (#2) was close to the median performance, ranking 13th out of 28 visual-based ones (MiAP, GMiAP) and 10th (F-ex). Finally, the multimodal runs ranked above average compared to competition. For instance, Run 3 ranked 15th out of the 35 multimodal runs (MiAP, GMiAP) and 18th (F-ex). Detailed results are accessible at the official task page.
 
Five runs were submitted. Run 1 ranked 5th out of 17 text-based methods (MiAP) and 7th (GMiAP, F-ex). The performance of the visual-based run (#2) was close to the median performance, ranking 13th out of 28 visual-based ones (MiAP, GMiAP) and 10th (F-ex). Finally, the multimodal runs ranked above average compared to competition. For instance, Run 3 ranked 15th out of the 35 multimodal runs (MiAP, GMiAP) and 18th (F-ex). Detailed results are accessible at the official task page.
Five runs were submitted. Run 1 ranked 5th out of 17 text-based methods (MiAP) and 7th (GMiAP, F-ex). The performance of the visual-based run (#2) was close to the median performance, ranking 13th out of 28 visual-based ones (MiAP, GMiAP) and 10th (F-ex). Finally, the multimodal runs ranked above average compared to competition. For instance, Run 3 ranked 15th out of the 35 multimodal runs (MiAP, GMiAP) and 18th (F-ex). Detailed results are accessible at the official task page.CERTH participated in ImageCLEF Photo Annotation and Retrieval 2012, and more specifically in the first subtask, namely Visual and Concept detection, annotation using Flickr photos. 
 
The objective of taking part in the competition was to evaluate two multimedia indexing approaches developed within SocialSensor and to compare the performance of methods using different sets of image features.
 
The first of the tested approaches constructs a similarity graph that includes both train and test images and trains concepts detectors using the graph Laplacian Eigenmaps (LE) as features. The second approach utilizes the concept of a “same class” model which takes as input the set of distances between the image to be annotated and a reference item representing a target concept, and predicts if the image belongs to the target concept.
 
Five runs were submitted. Run 1 ranked 5th out of 17 text-based methods (MiAP) and 7th (GMiAP, F-ex). The performance of the visual-based run (#2) was close to the median performance, ranking 13th out of 28 visual-based ones (MiAP, GMiAP) and 10th (F-ex). Finally, the multimodal runs ranked above average compared to competition. For instance, Run 3 ranked 15th out of the 35 multimodal runs (MiAP, GMiAP) and 18th (F-ex). Detailed results are accessible at the official task page.CERTH participated in ImageCLEF Photo Annotation and Retrieval 2012, and more specifically in the first subtask, namely Visual and Concept detection, annotation using Flickr photos. 
 
The objective of taking part in the competition was to evaluate two multimedia indexing approaches developed within SocialSensor and to compare the performance of methods using different sets of image features.
 
The first of the tested approaches constructs a similarity graph that includes both train and test images and trains concepts detectors using the graph Laplacian Eigenmaps (LE) as features. The second approach utilizes the concept of a “same class” model which takes as input the set of distances between the image to be annotated and a reference item representing a target concept, and predicts if the image belongs to the target concept.
 
Five runs were submitted. Run 1 ranked 5th out of 17 text-based methods (MiAP) and 7th (GMiAP, F-ex). The performance of the visual-based run (#2) was close to the median performance, ranking 13th out of 28 visual-based ones (MiAP, GMiAP) and 10th (F-ex). Finally, the multimodal runs ranked above average compared to competition. For instance, Run 3 ranked 15th out of the 35 multimodal runs (MiAP, GMiAP) and 18th (F-ex). Detailed results are accessible at the official task page.

Social media mining and multimedia analysis research and applications

Seminar at the Centre for Vision, Speech and Signal Processing (CVSSP), University of Surrey, "Social media mining and multimedia analysis research and applications". The presentation slides of the seminar are available here.

Syndicate content