Monday, July 26, 2010

ImageCLEF's Wikipedia Retrieval Task Results

ImageCLEF is the cross-language image retrieval track which is run as part of the Cross Language Evaluation Forum (CLEF) campaign. The ImageCLEF retrieval benchmark was established in 2003 with the aim of evaluating image retrieval from multilingual document collections. Images by their very nature are language independent, but often they are accompanied by texts semantically related to the image (e.g. textual captions or metadata). Images can then be retrieved using primitive features based on pixels with form the contents of an image (e.g. using a visual exemplar), abstracted features expressed through text or a combination of both. The language used to express the associated texts or textual queries should not a ect retrieval, i.e. an image with a caption written in English should be searchable in languages other than English.

ImageCLEF's Wikipedia Retrieval task provides a testbed for the systemoriented evaluation of visual information retrieval from a collection of Wikipedia images. The aim is to investigate retrieval approaches in the context of a large and heterogeneous collection of images (similar to those encountered on the Web) that are searched for by users with diverse information needs.
In 2010, ImageCLEF's Wikipedia Retrieval used a new collection of over 237000 Wikipedia images that cover diverse topics of interest. These images are associated with unstructured and noisy textual annotations in English, French, and German.

Based on the best run per group, as illustrated in the following Table, we are 2nd out of 13 in P@10:


Similarly, as illustrated in the following Table, based on the best run per group, we are 2nd out of 13 in P@20:


Overall, based on the best run per group, as illustrated in the following Table, we are 7th out of 13 in MAP:


More details about our fusion methods will be added soon

Friday, July 23, 2010

New Google Image Search Design Comes with New Ad Opportunity

Article From

Google began rolling out a new design for its image search today, and along with that came a new ad format. The format is called (appropriately) Image Search Ads.
"These ads appear only on Google Images, and they let you include a thumbnail image alongside your lines of text," explains Google Images Product Manager Nate Smith. "we hope they’re a useful way to reach folks who are specifically looking for images."
Advertisers can review specific performance metrics for their ads on Google Images. They are priced the same as standard AdWords ads with cost-per-click pricing.
Advertisers can create the ads using Google's Display Ad Builder. They can use a template to pair relevant ad text with targeted images.

Display Ad Builder

To create an Image Search Ad, simply go to your AdWords account, select the campaign or ad group where you want to create it, click the Ads tab, select Display ad builder from the "new ad" drop-down menu, then select "templates for search" and choose Image Search Ad.
Google recommends that advertisers create a new ad group so that they can target keywords, adjust bids, and track performance for ads specifically on Google Images.

ImageCLEF - Experimental Evaluation in Visual Information Retrieval

Series: The Information Retrieval Series, Vol. 32

Müller, H.; Clough, P.; Deselaers, Th.; Caputo, B. (Eds.)

Seven Years of Image Retrieval Evaluation

978-3-642-15180-4_Mueller_cover_RawData.indd The creation and consumption of content, especially visual content, is ingrained into our modern world. This book contains a collection of texts centered on the evaluation of image retrieval systems. To enable reproducible evaluation we must create standardized benchmarks and evaluation methodologies. The individual chapters in this book highlight major issues and challenges in evaluating image retrieval systems and describe various initiatives that provide researchers with the necessary evaluation resources. In particular they describe activities within ImageCLEF, an initiative to evaluate cross-language image retrieval systems which has been running as part of the Cross Language Evaluation Forum (CLEF) since 2003.

To this end, the editors collected contributions from a range of people: those involved directly with ImageCLEF, such as the organizers of specific image retrieval or annotation tasks; participants who have developed techniques to tackle the challenges set forth by the organizers; and people from industry and academia involved with image retrieval and evaluation generally.

Mostly written for researchers in academia and industry, the book stresses the importance of combing textual and visual information – a multimodal approach – for effective retrieval. It provides the reader with clear ideas about information retrieval and its evaluation in contexts and domains such as healthcare, robot vision, press photography, and the Web.

Read More

Tessera technology targets image enhancement and improvement

Article from digitalcamerareview

Tessera Technologies has developed video enhancement technology that is designed to compensate for camera shake and vibration, as well as improve overall digital image quality.

The FotoNation Video Tools is a hardware/software solution for camera manufacturers that combines video stabilization technology and image enhancement to help overcome the image quality problems inherent with small digital cameras and camcorders. The software can significantly improve video quality by removing or reducing the effect of jerks and frame jumps, while improving images by softening skin tones and removing blemishes that may become more apparent with high-definition video capture and playback.

Since the solution relies on both hardware and software to offload image stabilization and processing from a camera's core resources, Tessera claims it can also improve overall performance of the device and accelerate image improvement tasks. The solution can also improve battery life, since there is less demand on central resources, said a company spokesman.

The FotoNation Face Tools technology includes the FotoNation FaceTracker solution, plus the SmileCheck, BlinkCheck and Beautification extension modules. The Face Tools technology can automatically zero in on a subject's face and check focus, exposure and color balance. It also reduces the appearance of wrinkles or other skin imperfections for enhanced still image quality. The software can also automatically detect and remove red- and golden-eye defects.

Also included in the toolset is FotoNation Face Recognition technology, which can identify and tag faces for simplified image management, sharing and retrieval.

The FotoNation Video Tools is available for licensing by camera manufacturers. More information on the company's digital imaging solutions is available at the Tessera Web site.

Development of a DICOM Structured Report Viewer for IRMACON

For diagnosing in Radiology it can be very helpful to examine similiar pictures with regarding results, comparing them to the new one. This way a second opinion to the current results can be obtained. IRMA (Image Retrieval in Medical Applications) implements a similarity search based on a given PACS (Picture Archiving System). Within the scope of the project IRMACON, IRMA is integrated exemplarily into radiologic routine for any given CBIR-systems. The results of the similarity search are being processed into DICOM Structured Reporting Documents (SR) by an Evidence-Creator, realized in a diploma thesis. The DICOM SR contains information to source and resulting pictures as well as patient and device data. Picture data can also exist in form of Region of Interest (ROI) if similarities should be compared only for a specified area of the given picture. SRs are based on a defined template, which specifies the structure and value types.

jansen_en_detail[1] In the context of this student research project a web based viewer for SRs is to be realized, which allows displaying SR data, given in binary format, with all relevant information (source and resulting pictures, patient and device data, Regions of Interest). To be independent of any given PACS, the viewer contains several possibilities of displaying pictures, which are linked within the SR.

1) On the one hand, pictures can be directly loaded and displayed from the DICOM-PACS. Because DICOM pictures cannot be displayed directly by a webbrowser, they need to be converted into a compatible format. This is achieved by the use of the DICOM OFFIS Toolkit (DCMTK).

2) On the other hand, IRMA uses a matching database for the similarity search, which matches DICOM-IDs to corresponding IRMA-IDs (since the search works on PNG-files). So pictures, referenced with their DICOM-IDs from within the SR, can be loaded from the IRMA-DB. This way we do neither need to access the DICOM-DB nor need to convert the DICOM binary format. Though IRMA-DB access is done by the Evidence Creator.

The viewer is also to be entirely integrated into the IRMA framework and is based, to some amount, on its components. The whole project is developed in PHP so results can be generated dynamically during runtime. The single steps needed to display a SR are:

  1. Download SR from the DICOM-DB, with its unique DICOM-ID given as a parameter
  2. Converting the SR into a XML-file using OFFIS DCMTK
  3. Parsing the XML-file using Xerces
  4. Transforming the XML data into an HTML-file using PHP
  5. Downloading pictures as described above
  6. Displaying ROIs for given pictures, if available

Schedule: since 08/2009

arranged by: Sebastian Jansen

supervised by: Dipl.-Inform. Petra Welter

Read More

Thursday, July 22, 2010

Fourth ACM International Conference on Web Search and Data Mining (WSDM2011)

February 9-12, 2011, Hong Kong

Paper Submission Deadline: August 1, 2010
Workshop Proposal Deadline: July 25, 2010

WSDM (pronounced "wisdom") is the premier international ACM conference covering research in the areas of search and data mining on the Web. The 4th ACM WSDM Conference will take place in Hong Kong, during February 9-12, 2011.

WSDM publishes original, high quality papers and presentations related to search and data mining on the Web and the Social Web, with an emphasis on practical but principled novel models of search, retrieval and data mining, algorithm design and analysis, economics implications, and in-depth experimental analysis of accuracy and performance.

WSDM2011 will have keynotes, invited talks, full paper presentations, posters, tutorials, workshops, social events, etc.  Topics covered include but are not limited to:

Web Search

  * Web Dynamics and Search

  * Multifaceted and task-driven search

  * Multimedia Web search

  * Security and privacy in Web search and mining

  * User interfaces for search interaction

  * Distributed, meta, and peer-to-peer search

  * Vertical portals and search

  * Search quality benchmarking and evaluation

  * Ranking and machine learning for ranking

Web Mining

  * Web measurements, Web evolution, Web models

  * Clustering, classification, and summarization of Web data

  * Opinion mining and sentiment analysis

  * Multimodal data mining

  * Data, entity, event and relationship extraction

  * Sense and entity disambiguation

  * Data integration and data cleaning

  * Discovery-driven Web and social network mining

  * Traffic and log analysis

  * Algorithms and systems for Web-scale search and mining

Social Search and Mining

  * Social network analysis, theories, models, and applications

  * Social media analysis: blogs and friendship networks

  * Social reputation, influence, and trust

  * User profiling and recommendation systems

  * User activity modeling and exploitation

  * Personalized search and ranking

  * Tags, Users, and Search

Monday, July 19, 2010

ANN: large overhead camera person tracking dataset

We have been contructing a video dataset of detected people walking through the atrium of a building, where the data comes from an overhead camera.

Because of the data volume, we are not recording the video images, but are doing target detection on the images and save the detection records.

The records include a bounding box and a colour histogram for each detected target. Because of lighting, camera jitter and stationary objects, there are detections of non-people, too. After detection, we also have applied our person tracker, recording the trajectories as both lists of detections and splines.

By July 19, 2010, there were 29 million target detections, of which an estimated 8.3 million were real targets, resulting in 96K observed trajectories. And we're still acquiring data!

You might find it interesting to use the data. It's at:

The data was acquired in the Informatics Forum, the main building of the School of Informatics at the University of Edinburgh.

By the way, there are over 120 public image and video datasets on CVonline at:

Prof. Robert B. Fisher, School of Informatics, Univ. of Edinburgh

EECE@DUTH Robotics Team Papers

The Following papers have been accepted for presentation at the 1st International Conference for Undergraduate and Postgraduate Students in Computer Engineering, Informatics, related Technologies and Applications.

1.Designing and Development of a Puzzle Solver System Using Morphological Associative Memories

2[2] In this project a novel approach for solving two-dimensional puzzles and assembling the puzzle using the IR52C robotic arm is presented. When compared to many classical methods for resolving puzzles, the proposed method is characterized by low computational cost and ease of development. In order to achieve shape recognition, classical techniques from the field of image processing are combined with pattern association techniques. A camera captures an image of the whole disassembled puzzle, whose pieces are scattered in the working area of the robotic arm. In the first step, this image is converted to a binary form using Otsu’s method. Then, the connected components are identified and their inclination is corrected. In order to achieve recognition of every single shape, Morphological Associative Memories (ΜΑΜs) that have been previously trained are used. After object recognition, the system communicates with the robotic arm to assemble the pieces of the puzzle. Experimental results demonstrate the effectiveness of the proposed method.


2.A Novel Video Summarization Method Based on Compact Composite Descriptors and Fuzzy Classifier

3[1] In this paper, a novel method to generate video summaries is proposed, which is allocated mainly for being applied to on-line videos. The novelty of this approach lies in the fact that the authors of this paper transfer the video summarization problem to a single query image retrieval problem. This approach utilizes the recently proposed Compact Composite Descriptors (CCDs) and a fuzzy classifier. In particular, all the video frames are initially sorted according to the distance between an artificially generated, video depended, image. Then the ranking list is classified into a preset number of clusters using the Gustafson Kessel fuzzy classifier. The video abstract is calculated by extracting a representative key frame from every cluster. A significant characteristic of the proposed method is its ability to classify the frames of the video into one or more clusters. Experimental results are presented to indicate the effectiveness of the proposed approach.


3.Wired and Wireless Methods for Controlling a 5- DOF Robotic Arm

4[1] This paper presents an open-source Graphical User Interface (GUI) that works under the Windows OS (XP or later version) used for controlling a five-degrees-of-freedom robotic arm, including both a virtual remote controller and a command parser. We also provide a control method using a wireless natural-physics controller that translates human-hand movements to commands for the robotic arm. Authors of this paper are using modern console game
controllers that have built-in accelerometers and IR sensors. The resulting methods are tested on Eurobtec’ s IR52c.


4.An Open Source Set of Libraries to Control a 5-DOF Robotic Arm

1[2] This project introduces a new, open-source, lightweight software library that implements solutions to the Direct and Inverse kinematics problem for a 5-degrees-of-freedom robotic arm in general. In addition, a library for low level communication with the IR25c robotic arm has been developed, that can be used with the previously mentioned library and allows for easy manipulation of the robot. Both libraries are open source software, provided under the GNU General Public License and are designed with modularity in mind, so that they can be used together or separately in large projects. Download

Friday, July 16, 2010

Interesting Research Topics for MS Theses

Implementation and verification of low-level image descriptors for medical image retrieval

This work builds upon an existing framework for medical image retrieval developed as a course project at FAU (MEDIX) and expands it in two significant ways:

(1) it should incorporate additional low-level image descriptors recently proposed in the literature (e.g., CEDD, FCTH, etc.); and

(2) it should be evaluated against a widely used dataset of radiology images, e.g. IRMA (

The goal of this thesis is to implement and test 2-4 low-level image descriptors published in the literature, integrate them into MEDIX, and evaluate their performance against a standard dataset.

Supervisor: Dr Oge Marques

Department of Computer & Electrical Engineering and Computer Science (CEECS)

Florida Atlantic University (FAU)
777 Glades Road
Boca Raton, FL 33431 - USA

Read More


SMDT2010 || Second Workshop on Semantic Multimedia Database Technologies

I’ve accepted an invitation to become a PC Member of the 12th International Workshop of the Multimedia Metadata Community

The workshop is collocated with the 5th International Conference on Semantic and Digital Media Technologies (SAMT2010) in Saarbrücken, Germany - Dec 1-3, 2010.
Ontology-based systems have been developed to structure content and support knowledge retrieval and management. Semantic multimedia data processing and indexing in ontology- based systems is usually done in several steps. One starts by enriching multimedia metadata with additional semantic information (possibly obtained by methods for bridging the semantic gap). Then, in order to structure data, a localized and domain specific ontology becomes necessary since the data has to be interpreted domain-specifically. The annotations are stored in an ontology management system where they are kept for further processing. In this scope, Semantic Database Technologies are now applied to ensure reliable and secure access, efficient search, and effective storage and distribution for both multimedia metadata and data. Their services can be used to adapt multimedia to a given context based on multimedia metadata or even ontology information. Services automate cumbersome multimedia processing steps and enable ubiquitous intelligent adaptation. Both, database and automation support facilitate the ubiquitous use of multimedia in advanced applications.
Topics of interest are how multimedia metadata standards and multimedia ontologies are mapped or integrated into databases, how multimedia query languages are built, and how semantic queries are optimized and processed. Moreover, we are interested how multimedia data services are conceived to ensure interoperability, how to improve security and reliability of access and storage of multimedia data and metadata. Adaptation of multimedia, semantic enrichment of multimedia and bridging of media breaks are typical examples of advanced multimedia database services.
The workshop proceedings will be published at Selected papers will appear in a special issue of a journal. Previous editions of the workshop have selected for instance journals from the Springer Verlag.

Thursday, July 15, 2010

Is 'red' the same to all creatures?

Article from CNN

Violets are blue and roses are red, but maybe those colors are all in your head.

What does it mean for an object to be "red"? Is the way you perceive blueness the same as your neighbor? Your cousin? What about your dog?

Many scientists believe that humans have color vision that is generally consistent across populations and cultures, and that there are evolutionary reasons behind that constancy.

"Color vision is all about emotions and moods, and it has much deeper and richer connections to the rest of our perceptual worlds," said Mark Changizi, a cognitive scientist at Rensselaer Polytechnic Institute in Troy, New York.

Color vision in humans and animals

But some people really don't see the color red in the way that most do. About 8 percent of men have trouble differentiating between certain colors; less than 0.5 percent of women have this problem, according to the American Academy of Ophthalmology.

Color vision is based on photoreceptors in the eye called cones, of which there are about 6 million to 7 million in the human retina. Humans normally have three types of cones, corresponding to short, medium and long wavelengths of light. Purplish blues are at the short end, and reds are at the long end. They eye also has about 120 million rods, which detect light but not color.

According to some estimates, the human eye can distinguish about 1 million to 10 million different colors. A small minority of women actually have four kinds of cones in their eyes -- meaning they could theoretically see even more colors -- but only a genetic test can determine who has extra cones, and it's unclear exactly how differently they may see.

In most cases of colorblindness, the cone systems for either medium or long wavelengths do not work properly, resulting in reds, greens and perhaps yellows appearing very similar. But different people experience this to varying degrees. In rarer cases, people have trouble telling blue and yellow apart; the rarest of all make people see the world in grayscale.

Dogs and cats are generally colorblind, somewhat like humans who have trouble with reds and greens, but only see pale shades of color. On the other hand, they see better at night and have better peripheral vision. Insects see through photoreceptor units numbering in the hundreds or the thousands, almost like viewing the world as a mosaic. Some animals actually have better color vision than humans. Pigeons and goldfish, for example, can see ultraviolet light, which is invisible to people.

"We're great for mammals but pretty mediocre by broader standards," said David Hilbert, associate professor of philosophy at the University of Illinois at Chicago.

Biologists believe that animals' visual systems have evolved over millions of years and that the particular structures around today have persisted because they carried some survival benefit to the animals.

Read More

Wednesday, July 14, 2010

Compact Shape Portrayal Descriptor (CSPD)

The CSPD is a MPEG-like descriptor that contains conventional contour and region shape features with a wide applicability from any arbitrary shape to document retrieval through word spotting. It is also, compact (it less than 16 bytes) and requires low computational cost. The obtained retrieval results are satisfactory in comparison with other MPEG-7 shapes descriptors. Particular, it has been designed with attention to its size and storage requirements, keeping it as small as possible without limiting its discriminating ability.

More specifically, the Compact Shape Portrayal Descriptor is a forty one dimension vector that it is created from five distinct features that capture satisfactory the shape of an object. Also, the values of the vector are quantized for binary representation in three bits for each so its storage requirements are only 123 bits per shape. As the block diagram portrays, the computation of the CSPD can be easily parallize as each feature can be calculated separately.

The advantage of this descriptor is its balance between its size, its calculation time and its retrieval results. It is suitable for existing retrieval systems that they need to add shape information to their descriptors without compromising theirs storing and computational requirements.

The following journal paper describes at every detail this descriptor:

K. Zagoris, Ε. Kavallieratou and N. Papamarkos, «Image Retrieval Systems Based On Compact Shape Descriptor and Relevance Feedback Information», Journal of Visual Communication and Image Representation , DOI information:10.1016/j.jvcir.2011.03.002,2011.

It was integrated into a shape retrieval system located at The system is based on Microsoft .NET 3.5/WPF. The interface was build by the XALM language and for the program logic the C# was used. The new img(Anaktisi) web site (currently in developing) will integrate the CSPD, too.

You can download the CSPD DLL here (mirror). In addition to the CSPD DLL file, an example solution for the Visual Studio 2010 is provided, which it is an extension to the TSRD example solution. You can download the example solution from here (mirror). The following commands are only needed in order to calculate the CSPD of a image:

// Creating the CSPD object

// Load the Image from the pictureBox_word.
// Alternately you can load it from the hard disk.
// In the end you need a Bitmap object that it represents a
// black/white 24bit image.
Bitmap myImage = (Bitmap)pictureBox_word.Image.Clone();

// Get the Descriptor through the CSPD object
double[] myCSPDescriptor = myCSPD.GetCSPDescriptor(myImage);

For more information or questions email me at

Dr Konstantinos Zagoris ( received the Diploma in Electrical and Computer Engineering in 2003 from Democritus University of Thrace, Greece and his phD from the same univercity in 2010. His research interests include document image retrieval, color image processing and analysis, document analysis, pattern recognition, databases and operating systems. He is a member of the Technical Chamber of Greece.

Friday, July 9, 2010

3rd International Conference on Agents and Artificial Intelligence (ICAART)

The purpose of the 3rd International Conference on Agents and Artificial Intelligence (ICAART) is to bring together researchers, engineers and practitioners interested in the theory and applications in these areas. Two simultaneous but strongly related tracks will be held, covering both applications and current research work within the area of Agents, Multi-Agent Systems and Software Platforms, Distributed Problem Solving and Distributed AI in general, including web applications, on one hand, and within the area of non-distributed AI, including the more traditional areas such as Knowledge Representation, Planning, Learning, Scheduling, Perception and also not so traditional areas such as Reactive AI Systems, Evolutionary Computing and other aspects of Computational Intelligence and many other areas related to intelligent systems, on the other hand.
A substantial amount of research work is ongoing in these knowledge areas, in an attempt to discover appropriate theories and paradigms to use in real-world applications. Much of this important work is therefore theoretical in nature. However there is nothing as practical as a good theory, as Boltzman said many years ago, and some theories have indeed made their way into practice. Informatics applications are pervasive in many areas of Artificial Intelligence and Distributed AI, including Agents and Multi-Agent Systems; This conference intends to emphasize this connection, therefore, authors are invited to highlight the benefits of Information Technology (IT) in these areas. Ideas on how to solve problems using agents and artificial intelligence, both in R&D and industrial applications, are welcome. Papers describing advanced prototypes, systems, tools and techniques and general survey papers indicating future directions are also encouraged. Papers describing original work are invited in any of the areas listed below. Accepted papers, presented at the conference by one of the authors, will be published in the Proceedings of ICAART with an ISBN. A book with the best papers of the conferene will be published by Springer-Verlag. Acceptance will be based on quality, relevance and originality. Both full research reports and work-in-progress reports are welcome. There will be both oral and poster sessions.
Special sessions, dedicated to case-studies and commercial presentations, as well as tutorials dedicated to technical/scientific topics are also envisaged: companies interested in presenting their products/methodologies or researchers interested in holding a tutorial are invited to contact the conference secretariat. Additional information can be found at


  • Regular Paper Submission: July 26, 2010 (deadline extended)
  • Authors Notification (regular papers): October 06, 2010
  • Final Regular Paper Submission and Registration: October 26, 2010




A short list of presented papers will be selected
so that revised and extended versions of
these papers will be published by
Springer-Verlag in a CCIS Series book

A short list of papers presented at the
conference will be selected for publication of
extended and revised versions in a special issue of
JOPHA - Journal of Physical Agents

Monday, July 5, 2010

LEGO Mindstorms Sudoku Solver

Friday, July 2, 2010

Multimedia Grand Challenge!!!

ACM Multimedia 2010, October 25-29th 2010, Firenze, Italy

What problems do Google, Yahoo!, HP, Radvision, CeWe, Nokia and other companies see driving the future of multimedia? The Multimedia Grand Challenge is a set of problems and issues from these industry leaders, geared to engage the Multimedia research community towards solving relevant, interesting and challenging questions in the multimedia industry’s 2-5 year horizon. The Grand Challenge was first presented as part of ACM Multimedia 2009. and it will be presented again in slightly modified form at ACM Multimedia 2010. Researchers are encouraged to submit working systems in response to these challenges to win the grand Challenge competition!

We are happy to inform you that this year's Grand Challenge will award three prizes:

Gold medal 1500 USD

Silver medal 1000 USD

Bronze medal 500 USD


3DLife Sport Camera Networks

CeWe Photo Set Theme Identification

Google Personal Diaries

Google Video Genre Classification

HP Visual Communication

Nokia Photo Location & Orientation

Radvision Content Adaptation

Radvision VideoConf Experience

Yahoo! Novel Image Understanding

Yahoo! Video Segmentation

We encourage you to consider the challenges and submit your contribution to the ACM Multimedia 2010 Grand Challenge track. The top submissions will be presented in a special event during the ACM Multimedia 2010 conference in Florence Italy.

Based on these oral presentations in Florence, winners will be selected for Grand Challenge awards worth 1500 USD.

Paper Submission

New in 2010: _any_ long, short or demo paper submitted to the regular ACM Multimedia conference can be considered for the Grand Challenge. Regular papers that want to be considered should submit a 1-page additional statement indicating why this work addresses a certain challenge best. In addition, papers, demonstrations and ideas that do not fit the normal scientific style may be submitted to the Grand Challenge by submitting a separate 4-page paper to the Grand Challenge only. (We are afraid that we will NOT consider papers in this path that have been rejected for the scientific program.)

A 1-page statement to accompany an _accepted_ paper or a new 4-page grand challenge paper should be prepared in double-column ACM proceedings style and submitted via

The submissions must:

- Significantly address one of the industry challenges posted on the Grand Challenge web site.

- Depict working, presentable systems or demos.

- Describe why the system presents a novel and interesting solution.

A number of solutions (perhaps 10-20) will be selected as finalists and invited to describe their work, demonstrate their solution and argue for the paper's success in the Grand Challenge Session in Florence. Each finalist will have several minutes to present their case. Final winners will be chosen by industry scientists, engineers and business luminaries.

Key Dates

August 1, 2010 Submission deadline

September 1, 2010 Acceptance notification October 25, 2010 Conference starts

Thursday, July 1, 2010

3rd International Conference on SImilarity Search and APplications (SISAP 2010)

Paper submission: (NEW)
July 9th, 2010 July 2nd, 2010

The International Conference on Similarity Search and Applications (SISAP) is a conference devoted to similarity searching, with emphasis on metric space searching. It aims to fill in the gap left by the various scientific venues devoted to similarity searching in spaces with coordinates, by providing a common forum for theoreticians and practitioners around the problem of similarity searching in general spaces (metric and non-metric) or using distance-based (as opposed to coordinate-based) techniques in general.

SISAP aims to become an ideal forum to exchange real-world, challenging and exciting examples of applications, new indexing techniques, common testbeds and benchmarks, source code, and up-to-date literature through a web page serving the similarity searching community. Authors are expected to use the testbeds and code from the SISAP web site for comparing new applications, databases, indexes and algorithms.

After the very successful first events in Cancun, Mexico (2008) and Prague, Czech Republic (2009), this year SISAP conference will be held in Istanbul, Turkey, on September 18-19, 2010.

The four best papers will be invited to be published in a special issue of Journal of Discrete Algorithms (Elsevier). SISAP 2010 is organized in cooperation with ACM SIGSPATIAL and the papers will be indexed in the ACM Digital Library.

World Cup VS PhD

MMRetrieval.NET (Alpha)

In this web application, we present an experimental multimodal search engine, which allows multimedia and multi-language queries, and makes use of the total available information in a multimodal collection. All modalities are indexed and searched separately, and results can be fused with different methods depending on

  • the noise and completeness characteristics of the modalities in a collection
  • whether the user is in a need of high initial precision or high recall

Beyond fusion, we also provide 2-stage retrieval by first thresholding the results obtained by secondary modalities targeting recall, and then re-ranking them based on the primary modality. The engine demonstrates the feasibility of the proposed architecture and methods on the ImageCLEF 2010 Wikipedia collection. Its primary modality is image, consisting of 237434 items, associated with noisy and incomplete user-supplied annotations and the Wikipedia articles containing the images. Associated modalities are written in any combination of English, German, French, or any other undetermined language.