Tuesday, November 30, 2010

Research Fellow - Image Retrieval

Research Fellow - Image Retrieval

Centre for Interactive Systems Research, City University London

This post offers the opportunity to work on an EPSRC-TSB funded project Piclet. The project will create a platform for image retrieval, share, use, and recommendation. It will provide users with an environment for interactive image retrieval with possibilities for buying and selling images. The project considers the image information seeking and retrieval behaviour of different user groups including professionals. Partners from industry include an SEO (search engine optimisation) company and a media advertising company.

In this post you will develop novel image retrieval algorithms based on content-based techniques, text retrieval, and user's context. You will implement and integrate these in an image retrieval system. The project will make use of current state-of-art image retrieval techniques, collections, relevant open-source initiatives, and evaluation methods. You will be involved in integrating these, creating user interfaces, and developing new system functionality where necessary.

Applicants should be qualified in Information Retrieval (or related area).

More specifically, text and image retrieval techniques will be needed. The qualification sought is PhD or equivalent experience in field. Strong programming skills are essential, and experience of any of the following areas is an advantage: Information Retrieval (IR); User-centred IR system design, implementation, and evaluation; Context learning, Adaptive IR, Recommender systems, Social media; User Interface design and evaluation.

Experience in collaborating with partners from industry is desirable.

For informal enquiries, please contact Dr. Ayse Goker:

Closing Date: 10 December, 2010

Details available via (The post will also be advertised on )

Monday, November 29, 2010

Public Thesis Defense–Invitation

My thesis has officially passed committee and, after a final check-the-commas read-through, will be on its way to the external examiners.

December 09, 10:00, Democritus University of Thrace

PhD_Cover - Final

If you are, or will be, in Xanthi, it would be great if you could join in person. If you are not, there will be a Skype conference call, with preference given to research participants who would like to listen in. Please email me to reserve a slot.

Abstract – Introduction

This chapter lists the goals and the contribution of the current thesis. The goals which were set up at the beginning of this work and which were adjusted during the process are:

  • The creation of a new family of descriptors which will combine more than one low levels feature in a compact vector, and which will have the ability to be incorporated in the pre-existing  MPEG-7 standard. The descriptors will be constructed via intelligent techniques.
  • The creation of a method for accelerating the searching procedure.
  • The investigation of several Late Fusion methods for image retrieval.
  • The creation of methods which will allow the use of the proposed descriptors in distributed image databases.
  • The development of a software which will contain a great amount of  descriptors proposed in the literature.
  • The development of open source libraries which will utilize the proposed descriptors as well as the MPEG-7 descriptors.
  • The creation of a new method for  encrypting  images  which will utilize features and parameters from the image retrieval field.
  • The creation of a new method and system implementation which will employ the proposed descriptors in order to achieve video summarization.
  • The creation of a new method and system implementation for image retrieval based on``Keywords'' which will be automatically generated via the use of the proposed descriptors.
  • Finally, the creation of a new method and system implementation for multi-modal search. The system will utilize both low level elements(which will originate from the proposed descriptors) as well as  high level elements (which will originate from keywords which will accompany the images).                            

Thesis Details
In the past few years there has been a rapid increase in the field of multi-media data, mostly due to the evolution of information technology. One of the main components of multi-media data is that of visual multi-media data, which includes digital images and video. While the issue of producing, compressing and propagating such media might have been a subject of scientific interest for a long time, in the past few years, exactly due to the increase in the range of data, a large part of the research was turned towards the management of retrieval of such materials.

The first steps in automated management and retrieval of visual multi-media, can be traced back to 1992, where the term Content Based Retrieval}was initially used. Since then, a new research field was created, which, approximately 20 years later, still remains active. And while initially this field of research seemed to be a research element classified under the general spectrum of information retrieval, as the years progressed, this research objective, has managed to attract scientists from various disciplines.

Even though there are a large number of scientists which occupy themselves with this field, no satisfactory and widely accredited solution to the problem has been proposed. The second Chapter of this thesis outlines a brief overview of the Fundamentals of Content-Based Image Retrieval.

During the course of this thesis, a study carried out that describes the most commonly used methods for retrieval evaluation and notes their weaknesses. It also proposes a new method of measuring the performance of retrieval systems and an extension of this method so that during the evaluation of retrieval results the parameters describing both the size of the database in
which the search is being executed as well as the size of the ground truth of each query are taken into account. The proposed method is generic and can be used for evaluating the retrieval performance of any type of information. This work is described in details in Chapter 3.

The core of the method proposed in this thesis is incorporated into the second thematic unit. This section includes a number of low level descriptors, whose features originate from the content of multi-media data which they describe. In contrast to  MPEG-7, each type of multi-media data will be described by a specific group of descriptors. The type of material will be determined by the content it describes. The descriptors created originate from fuzzy methods and are characterized by their low storage requirements (23-72 bytes per image).  Moreover, each descriptor combines the structure of more than one features (ie color and texture). This attribute classifies them as composite descriptors. The sum of descriptors which are incorporated into the second thematical unit of the thesis can be described by the general term  Compact Composite Descriptors.

In its entirety, the second thematic unit of the thesis contains descriptors for the following types of multi-media material:

  • Category 1: Images/ Video with natural content
  • Category 2: Images/ Video with artificially generated content
  • Category 3: Images with medical content

For the description and retrieval of multi-media material with natural content, 4 descriptors were developed:

  • CEDD - Color and Edge Directivity Descriptor
  • C.CEDD - Compact Color and Edge Directivity Descriptor
  • FCTH    - Fuzzy Color and Texture Histogram
  • C.FCTH - Compact Fuzzy Color and Texture Histogram

The CEDD includes texture information produced by the six-bin histogram of a fuzzy system that uses the five digital filters proposed by the MPEG-7 EHD. Additionally, for color information the CEDD uses a 24-bin color histogram produced by the 24-bin fuzzy-linking system. Overall, the final histogram has 6 X 24=144 regions.

The FCTH descriptor includes the texture information produced in the eight-bin histogram of a fuzzy system that uses the high frequency bands of the Haar wavelet transform. For color information, the descriptor uses a 24-bin color histogram produced by the 24-bin fuzzy-linking system. Overall, the final histogram includes 8 X 24=192 regions.

The method for producing the C.CEDD differs from the CEDD method only in the color unit. The C.CEDD uses a fuzzy ten-bin linking system instead of the fuzzy 24-bin linking system. Overall, the final histogram has only 6 X 10=60 regions. Compact CEDD is the smallest descriptor of the proposed set requiring less than 23 bytes per image.


The method for producing C.FCTH differs from the FCTH method only in the color unit. Like its C.CEDD counterpart, this descriptor uses only a fuzzy ten-bin linking system instead of the fuzzy 24-bin linking system. Overall, the final histogram includes only 8 X 10=80 regions.

To restrict the proposed descriptors' length, the normalized bin values of the descriptors are quantized for binary representation in a three bits/bin quantization.

Experiments conducted on several benchmarking image databases demonstrate the effectiveness of the proposed descriptors in outperforming the MPEG-7 Descriptors as well as other state-of-the-art descriptors from the literature. These descriptors are described in details in Chapter 5.

Chapter 6 describes the Spatial Color Distribution Descriptor (SpCD). This descriptor combines
color and spatial color distribution information. Since these descriptors capture the layout information of color features, they can be used for image retrieval by using hand-drawn sketch queries. In addition, the descriptors of this structure are considered to be suitable for colored graphics, since such images contain relatively small number of color and less texture regions than the natural color images. This descriptor uses a new fuzzy-linking system, that maps the colors of the image in a custom 8 colors palette.

The rapid advances made in the field of radiology, the increased frequency in which oncological diseases appear, as well as the demand for regular medical checks, led to the creation of a large database of radiology images in every hospital or medical center. There is now the imperative need to create an effective method for the indexing and retrieval of these images. Chapter 7 describes a new method of content based radiology medical image retrieval using the Brightness and Texture Directionality Histogram (BTDH). This descriptor uses brightness and texture characteristics as well as the spatial distribution of these characteristics in one compact 1D vector. The most important characteristic of the proposed descriptor is that its size adapts according to the storage capabilities of the application that is using it.

The requirements of the modern retrieval systems are not limited to the achievement of good retrieval results, but extend to their ability for quick results. The majority of the Internet users would accept a reduction in the accuracy of the results in order to save time from searching.  The third thematic unit describes how the proposed descriptors may be modified, in order to achieve a faster retrieval from databases. Test results indicate that the developed descriptors are in a position to execute retrieval of approximately 100,000 images per second, regardless of dimensions. Details on the method developed are given in Chapter 8.

In Chapter 9 the procedure of early fusion of the two descriptors which describe visual multi-media material with natural content, is described. Given the fact that this category includes more than one descriptors, the procedure for combining these descriptors in order to further improve on the retrieval results, is analyzed.

The proposed descriptors are capable of describing images with a specific content. The  descriptors developed for use with images with natural content cannot be used to retrieve grayscale medical images and vice versa. Due to this, the calculation of the efficiency of each descriptor was employed using image databases with homogenous content, suitable for the specific descriptor. However, the databases mostly used in the Internet are heterogeneous, and include images from every category. The fourth thematic unit of this thesis describes how late fusion techniques can be used to combine all the proposed descriptors, in order to achieve high retrievals scores in databases of this kind. Linear and non linear methods, which were adopted from the information retrieval field, have proven that the combination of descriptors yields very satisfactory results when used in heterogeneous data bases.

In the same, fourth, thematic unit a retrieval scenario from distributed image databases is considered. In this scenario, the user executes a search in multiple databases. However, it is possible that each database uses its own descriptor(s) for the images it contains. Adopting once more methods from the information retrieval field and combining them with a method developed in this thesis it is possible to achieve high retrieval scores. Details on the fusion methods, as well as the retrieval methods from distributed image databases, are given in Chapter 10.

Finally, the fourth thematic unit is completed by a relevance feedback algorithm. The aim of the proposed algorithm is to better readjust or even to alter the initial results of the retrieval, based on user preferences. During this process, the user selects from the retrieved results those images which are similar to his/her expected results. Information extracted from these images is in the sequel used to alter the descriptor of the query image. The method is described in Chapter 11.

The fifth  part of the thesis describes the implementation of the four prior parts into free and open source software packages. During the course of the thesis, 3 software packages were developed:

  • Software 1: Img(Rummager)
  • Software 2: Img(Anaktisi)
  • Software 3: LIRe

Img(Rummager) was employed for the demonstration of the results of the research carried out in this thesis. In addition to the developed descriptors, the software implements a large number of descriptors from the literature (including the MPEG-7 descriptors), so that the application constitutes a platform for retrieving images via which the behavior of a number of descriptors can be studied. The application can evaluate the retrieval results, both via the use of the new image retrieval evaluation method as well as via  MAP and ANMRR. The application was programmed using  C# and is freely available via the  ``Automatic Control, Systems and Robotics Laboratory'' webpage, Department of Electrical and Computer Engineering, Democritus University of Thrace.

Img(Anaktisi) was developed in collaboration with``Electrical Circuit Analysis Laboratory'', Department of Electrical and Computer Engineering, Democritus University of Thrace, and is an Internet based application which possesses the capability of executing image retrieval using the proposed in these thesis tools in a large number of images. The application is programmed in C\# .

Moreover, the proposed descriptors were included into the  free and open source library, LiRE. This library is programmed in JAVA and includes implementations for the most important  descriptors used for image retrieval. The program was developed in collaboration with the ALPEN-ADRIA University of Information Technology in Klagenfurtm Austria, Distributed Multimedia Systems Research Group (Ass. Prof. M. Lux). Details on the developed software are given in Chapter 12.

The sixth thematical unit of the thesis presents some of the applications which were developed via the use of the method which was employed during the thesis research. Initially, a system of image encryption was developed, which adopts methods from the  field of image retrieval. The proposed method employs cellular automata and image descriptors in order to ensure the safe transfer of images and video. Chapter 13 describes the method.

In collaboration with ``Electrical Circuit Analysis Laboratory'', Department of Electrical and Computer Engineering, Democritus University of Thrace, an automated image annotation system was developed, using support vector machines. The combination of descriptors characterizes the image content with one or more words from a predetermined dictionary. Both the developed system, as well as the details regarding the method are given in Chapter 14.

In addition, a system which combines all the descriptors from the second thematical unit, as well as the fusion methods of the fourth thematic unit, was developed in order to create automated video summaries. The method utilized fuzzy classificators in order to create a summary in a predetermined class number, with the unique attribute of multiple participation of each frame in each class. Details are given in Chapter 15.

For the purposes of this thesis, an application was developed, which combines the proposed descriptors with high level features. Specifically, an application was developed which combines the visual content of  250,000 images from the Wikipedia, with the tags which accompany the images, as well as with the content of articles in which these images can be found. In reality, this problem is a fusion problem with multiple modalities and is described in Chapter 16.

The  MPEG-7 standard proposed a structure via which visual-acoustic multi-media data bases are described. Each database is described via an XML file which contains the information for each image for a standardized format. This structure allows other applications to expand their structure by adding new fields. The first appendix of the thesis analyzes how the developed descriptors can be incorporated into the standardized MPEG-7 format.

Saturday, November 27, 2010

Hessian Optimal Design for image retrieval

Ke Lu, Jidong Zhao and Yue Wu

Available online 19 November 2010.


Recently there has been considerable interest in active learning from the perspective of optimal experimental design (OED). OED selects the most informative samples to minimize the covariance matrix of the parameters so that the expected prediction error of the parameters, as well as the model output, can be minimized. Most of the existing OED methods are based on either linear regression or Laplacian regularized least squares (LapRLS) models. Although LapRLS has shown better performance than linear regression, it suffers from the fact that the solution is biased towards a constant and the lack of extrapolating power. In this paper, we propose a novel active learning algorithm called Hessian Optimal Design (HOD). HOD is based on the second-order Hessian energy for semi-supervised regression which overcomes the drawbacks of Laplacian based methods. Specifically, HOD selects those samples which minimize the parameter covariance matrix of the Hessian regularized regression model. The experimental results on content-based image retrieval have demonstrated the effectiveness of our proposed approach.

Keywords: Active learning; Regularization; Image retrieval

Read More

Introduction to Content Based Image Retrieval Tutorial

Introduction to Content Based Image Retrieval by Bin Shen


Available at

Tuesday, November 23, 2010

20 Must-See TED Speeches for Computer Scientists

Article from

TED is a nonprofit organization that has dedicated itself to the concept that good ideas are worth spreading. To promote these great ideas, TED (which stands for Technology, Entertainment, Design) has a yearly conference in Long Beach, California. While the conference is too expensive for most people to attend, at $6000 per person, TED has used the Internet to share talks from the conference since 2006.

These are free of charge to the public and an amazing resource for professionals in many fields. Computer scientists will find these 20 TED speeches (not ranked in any particular order) informative, challenging, and stimulating. Maybe an idea worth spreading will help you make the connection you’ve been looking for in your own computer science research.

    1. George Dyson at the birth of the computer | Video on

    Have you ever heard the real stories about the dawn of computing? Sure, you have probably memorized important dates and can talk about things like vacuum tubes and punch cards, but do you really know what went on at the beginning? In this enlightening talk, computer historian George Dyson reveals fascinating and sometimes funny anecdotes about the beginning of computing.

    2. Kwabena Boahen on a computer that works like the brain | Video on

    Kwabena Boahen’s team at Stanford is working on computer technology that works less like a computer usually does, and more like the human brain does. He discusses the inefficiencies associated with traditional computing and ways that they might be overcome using reverse engineering of the human nervous system.

    3. Jeff Han demos his breakthrough touchscreen | Video on

    Jeff Han’s incredible speech from 2006 shows the future of touch-screen interfaces. His scalable, multi-touch, pressure-sensitive interface allows people to use the computer without the barriers of having to point and click all the time.

    4. Paul Debevec animates a photo-real digital face | Video on

    Paul Debevec explains the process behind his animation of Digital Emily, a hyper-realistic character. Digital Emily is based on a real person named Emily, and created using an advanced 360-degree camera setup.


    In this mind-boggling speech, Stephen Wolfram, a world-renowned leader in scientific computing, discusses his life’s mission. He wants to make all of the knowledge in the world computational. His new search engine, Wolfram|Alpha, is designed to take all of the available information on the web and make instant computations of that information accessible to the world.

    6. Dennis Hong: My seven species of robot

    As artificial intelligence technology continues to evolve, the quest continues to create robots that are truly useful on a scientific and on an ordinary, day-to-day level. Dennis Hong’s RoMeLa lab at Virginia Tech presents seven distinctly different robots in this talk. Make sure you watch all the way to the end to discover his secrets of creativity.


    Gary Flake believes that the whole of all data we have is greater than the sum of its parts. He and the staff at Microsoft Live Labs have created a fascinating new tool called Pivot that enables people to browse the web not by going from page to page, but by looking at large patterns all at once.

Read More

Monday, November 15, 2010

Relative status of journal and conference publications in computer science

Though computer scientists agree that conference publications enjoy greater status in computer science than in other disciplines, there is little quantitative evidence to support this view. The importance of journal publication in academic promotion makes it a highly personal issue, since focusing exclusively on journal papers misses many significant papers published by CS conferences.

Here, we aim to quantify the relative importance of CS journal and conference papers, showing that CS papers in leading conferences match the impact of papers in mid-ranking journals and surpass the impact of papers in journals in the bottom half of the Thompson Reuters rankings ( for impact measured in terms of citations in Google Scholar. We also show that poor correlation between this measure and conference acceptance rates indicates conference publication is an inefficient market where venues equally challenging in terms of rejection rates offer quite different returns in terms of citations.

How to measure the quality of academic research and performance of particular researchers has always involved debate. Many CS researchers feel that performance assessment is an exercise in futility, in part because academic research cannot be boiled down to a set of simple performance metrics, and any attempt to introduce them would expose the entire research enterprise to manipulation and gaming. On the other hand, many researchers want some reasonable way to evaluate academic performance, arguing that even an imperfect system sheds light on research quality, helping funding agencies and tenure committees make more informed decisions.

One long-standing way of evaluating academic performance is through publication output. Best practice for academics is to write key research contributions as scholarly articles for submission to relevant journals and conferences; the peer-review model has stood the test of time in determining the quality of accepted articles. However, today's culture of academic publication accommodates a range of publication opportunities yielding a continuum of quality, with a significant gap between the lower and upper reaches of the continuum; for example, journal papers are routinely viewed as superior to conference papers, which are generally considered superior to papers at workshops and local symposia. Several techniques are used for evaluating publications and publication outlets, mostly targeting journals. For example, Thompson Reuters (the Institute for Scientific Information) and other such organizations record and assess the number of citations accumulated by leading journals (and some high-ranking conferences) in the ISI Web of Knowledge ( to compute the impact factor of a journal as a measure of its ability to attract citations. Less-reliable indicators of publication quality are also available for judging conference quality; for example, a conference's rejection rate is often cited as a quality indicator on the grounds that a high rejection rate means a more selective review process able to generate higher-quality papers. However, as the devil is in the details, the details in this case vary among academic disciplines and subdisciplines.

Here, we examine the issue of publication quality from a CS/engineering perspective, describing how related publication practices differ from those of other disciplines, in that CS/engineering research is mainly published in conferences rather than in journals. This culture presents an important challenge when evaluating CS research because traditional impact metrics are better suited to evaluating journal rather than conference publications.

In order to legitimize the role of conference papers to the wider scientific community, we offer an impact measure based on an analysis of Google Scholar citation data suited to CS conferences. We validate this new measure with a large-scale experiment covering 8,764 conference and journal papers to demonstrate a strong correlation between traditional journal impact and our new citation score. The results highlight how leading conferences compare favorably to mid-ranking journals, surpassing the impact of journals in the bottom half of the traditional ISI Web of Knowledge ranking. We also discuss a number of interesting anomalies in the CS conference circuit, highlighting how conferences with similar rejection rates (the traditional way of evaluating conferences) can attract quite different citation counts. We also note interesting geographical distinctions in this regard, particularly with respect to European and U.S. conferences.

author = {Freyne, Jill and Coyle, Lorcan and Smyth, Barry and Cunningham, Padraig},
title = {Relative status of journal and conference publications in computer science},
journal = {Commun. ACM},
volume = {53},
issue = {11},
month = {November},
year = {2010},
issn = {0001-0782},
pages = {124--132},
numpages = {9},
url = {},
doi = {},
acmid = {1839701},
publisher = {ACM},
address = {New York, NY, USA},

Sunday, November 14, 2010

Elsevier Releases Image Search; New SciVerse ScienceDirect Feature

Elsevier a leading publisher of scientific, technical and medical (STM) information, today announced the availability of Image Search, a new SciVerse ScienceDirect ( feature that enables users to quickly and efficiently find images and figures relevant to their specific research objectives.

The new feature allows researchers to search across more than 15 million images contained within SciVerse ScienceDirect. Results include tables, photos, figures, graphs and videos from trusted peer-reviewed full text sources. Saving researchers significant time, Image Search ensures increased efficiency in finding the relevant visuals.

Search results can be refined by image type and contain links to the location within the original source article, allowing researchers to verify image in context. Researchers can take advantage of Image Search to learn new concepts, prepare manuscripts or visually convey ideas in presentations and lectures. the new feature is available to SciVerse ScienceDirect subscribed
users at no additional cost.

“By allowing researchers to delve immediately into the images on SciVerse ScienceDirect, the new feature saves researchers valuable time and offers immediate access to a vast database of trusted visual content,” said Rafael Sidi, Vice President of Product Management for Elsevier’s SciVerse ScienceDirect. “Developed in response to user feedback, Image Search represents another example of the steps Elsevier is taking to speed scientific search and discovery by improving researcher workflow.”

Saturday, November 13, 2010

All-seeing eye for CCTV surveillance

Reported by Duncan Graham-Rowe 09 November 2010 in New Scientist

In some cities, CCTV is a victim of its own success: the number of cameras means there is not enough time to review the tapes. That could change, with software that can automatically summarise the footage.

"We take different events happening at different times and show them together," says Shmuel Peleg of BriefCam based in Neve Ilan, Israel, a company he co-founded based on his work at the Hebrew University of Jerusalem. "On average an hour is summarised down to a minute."

The software relies on the fact that most surveillance cameras are stationary and so record a static background. This makes it possible for the software to detect and extract sections of the video when anything moving enters the scene. All events occurring in a given time period, say an hour, are then superimposed to create a short video that shows all of the action at once (see video).

If something of interest is spotted while watching the summary, the operator can click on it to jump straight to the relevant point in the original CCTV footage. BriefCam's software has been shortlisted as a finalist for the 2010 Global Security Challenge prize at its conference in London this week.

Thursday, November 11, 2010


CAIP2011 is the fourteenth in the CAIP series of biennial international conferences devoted to all aspects of Computer Vision, Image Analysis and Processing, Pattern recognition and related fields. CAIP2011 will be held in August 29-31, 2011, in Seville (Spain). The last CAIP conferences were held in Groningen (2003), Paris (2005), Vienna (2007) and Münster (2009). CAIP2011 will be hosted by Seville University. The scientific program of the conference will consist of several keynote addresses, high quality papers selected by the international program committee and presented in a single track. Poster presentations will allow expert discussions of specialized research topics.

he scope of CAIP'11 includes, but not limited to, the following areas:

  • 3D Vision
  • 3D TV
  • Biometrics
  • Color and texture
  • Document analysis
  • Graph-based Methods
  • Image and video indexing and database retrieval
  • Image and video processing
  • Image-based modeling
  • Kernel methods
  • Medical imaging
  • Mobile multimedia
  • Model-based vision approaches
  • Motion Analysis
  • Non-photorealistic animation and modeling
  • Object recognition
  • Performance evaluation
  • Segmentation and grouping
  • Shape representation and analysis
  • Structural pattern recognition
  • Tracking
  • Applications


The conference proceedings will be published in the Springer LNCS series.
A special issue with selected papers in Pattern Recognition will be published in Pattern Recognition Letters.
A special issue with selected papers in Analysis of Images will be published in Journal of Mathematical Imaging and Vision .

Monday, November 8, 2010

MediaEval Benchmarking Initiative

MediaEval LogoMediaEval is a benchmarking initiative dedicated to evaluating new algorithms for multimedia access and retrieval. MediaEval focuses on speech, language and contextual aspects of video (geographical and social context). Typical tasks include, predicting tags for video, both user-assigned tags (Tagging Task) and geo-tags (Placing Task). We are daring enough to tackle "subjective" aspects of video (Affect Task).

How can I get involved?
MediaEval is an open initiative, meaning that any interested research group is free to signup and participate. Signup for MediaEval 2011 will open in early Spring 2011 on this website. Groups sign up for one or more tasks, they then receive task definitions, data sets and supporting resources, which they use to develop their algorithms. At the end of the summer, groups submit their results and in the fall they attend the MediaEval workshop. More information on MediaEval is available on theFAQ page (see also Why Participate?)

Which tasks will be offered in 2011?
The task offering in 2011 will be finalized on the basis of participant interest, assessed, as last year, via a survey. At this time, we anticipate that we will run a Tagging Task and Placing Task (see MediaEval 2010 for descriptions of the form these tasks took in 2010). We also anticipate offering an additional task involving ad hoc video retrieval and one involving violence detection. Please watch this page for additional information to appear at the beginning of 2011 or contact the organizers for information on the current status of plans.

Beyond plain sight?
We are interested in pushing video retrieval research beyond the pictorial content of video. In other words, we want to get past what we see in "plain sight" when watching a video and instead also analyze video content in terms of its overall subject matter (what it is actually about) and the effect that it has on viewers. This goal is consistent with our emphasis on speech, language and context features (geographical and social context).

There's a MediaEval trailer?
Yes. In the MediaEval 2010 Affect Task, we analyzed a travelogue series created by Bill Bowles, filmmaker and traveling video blogger. We were looking for elements that could be used to automatically predict whether, overall, a general audience would find a particular episode engaging or boring. Bill attended the MediaEval 2010 workshop in Pisa, listened to our findings and created a travelogue episode for us. In this video, Bill tells the story of MediaEval in his own words...and makes conscious use of some of the predictors we found to correlate with viewer engagement.

Wednesday, November 3, 2010

Facial recognition security to keep handset data safe

Reported By Stewart Mitchell in PCPro – Article from

Mobile phone security could come face-to-face with science fiction following a demonstration of biometric technology by scientists at the University of Manchester.

Mobile phones carry an ever-increasing level of personal data, but PIN-code log-ins can prove weak – if even applied – leaving those details vulnerable if a handset is stolen.

Facial recognition software could change that picture drastically.

“The idea is to recognise you as the user, and it does that by first taking a video of you, so it has your voice and lots of images for comparison,” said Phil Tresadern, lead researcher on the project.

Before letting anyone access the handset, the software examines an image of the person holding the handset, cross-referencing 22 geographical landmarks on the face with that of the owner.

We had to change some of the floating point calculations to fixed points with whole numbers to make it more efficient, but it ported across quite easily

“Existing mobile face trackers give only an approximate position and scale of the face,” said Tresadern. “Our model runs in real time and accurately tracks a number of landmarks on and around the face such as the eyes, nose, mouth and jaw line.”

The scientists claim their method is unrivalled for speed and accuracy and works on standard smartphones with front-facing cameras.

Facial recognition technology is nothing new, but squeezing something so computationally complicated onto a smartphone is quite an achievement.

“We have a demo that we have shown to potential partners that is running on Linux on a Nokia N900,” said Tresadern, adding that his team had to streamline the software to make it run on a mobile phone.

“We had to change some of the floating point calculations to fixed points with whole numbers to make it more efficient,” he said. “But other than that it ported across quite easily.”

The Manchester team said the technology has already attracted interest from “someone interested in putting it into the operating system for Nokia” and that the software could be licensed by app developers for any mobile device.
Read more: Facial recognition security to keep handset data safe | Security | News | PC Pro

Tuesday, November 2, 2010

The IASTED International Conference on Signal and Image Processing and Applications ~SIPA 2011~


Prof. Ioannis Andreadis
Democritus University of Thrace, Greece


Situated in the warm and sunny Mediterranean, Crete is the largest of the Greek islands. It is renowned in myth and rich in a history that spans thousands of years. The centre of the ancient Minoan civilisation, Crete is the backdrop to the dramatic legend of the Minotaur. Crete's history is intertwined with struggles for regional dominance; over centuries, Crete has been ruled by different peoples, including the Venetians, the Ottoman Turks, the Romans, and the Byzantines, all of whom of have left their mark on the island.
Today, history and myth blend seamlessly with Crete's natural beauty and lively culture. Take in the sharp mountain ranges, interrupted by steep ravines and divided by fields of olive trees, all below a splendid sky. Spend a day exploring the wild cypress forests and keep an eye out for the kri-kri, the Cretan wild goat. A trip to the stunning White Mountains and the Samariá Gorge is a must, as are the many ruins and historic landmarks throughout the island. Unwind with a sunset stroll on the beach followed by a hearty meal at a local tavern, or an evening of vibrant city nightlife.


The IASTED International Conference on Signal and Image Processing and Applications (SIPA 2011) will be an international forum for researchers and practitioners interested in the advances in and applications of signal and image processing. It is an opportunity to present and observe the latest research, results, and ideas in these areas. SIPA 2011 aims to strengthen relationships between companies, research laboratories, and universities. All papers submitted to this conference will be double blind evaluated by at least two reviewers. Acceptance will be based primarily on originality and contribution.

Monday, November 1, 2010

MMRetrieval Presentation

Visit MMretrieval Experimental multimodal search engine

Orasis (ver. 1.1)

Orasis Image Processing Software (ver. 1.1) is now freely available for download at:

You can improve you photos, easily, with just a few mouse clicks.

Orasis is an experimental, biologically-inspired, image enhancement software, which employs the characteristics of the center-surround cells of the Human Visual System.

Many times the image captured by a camera and the image in our eyes are dramatically different. Especially when there are shadows or highlights in the same scene. In these cases our eyes can distinguish many details in the shadows or the highlights, while the image captured by the camera suffers from loss of visual information in these regions.
Orasis attempts to bridge the gap between "what you see" and "what the camera outputs". It enhances the shadow or the highlight regions of an image, while keeping intact all the correctly exposed ones. The final result is a lot closer to the human perception of the scene, than the original captured image, revealing visual information that otherwise wouldn't be available to the human observer. Additionally to the above, Orasis can correct low local contract, colors and noise.

The new version includes:

· JPEG compression support

· EXIF data support

· Keyboard shortcuts

· Decorrelation between brightness and local contrast enhancement

· Automatic Color Saturation

· Automatic check for Updates

· Improved memory management

· Improved Noise reduction algorithms

· Improved Automatic selection of parameters

· Improved Color Correction algorithm

· Improved Independent RGB mode

· Improved Interface

(Greek Version)

Η νέα έκδοση του λογισμικού Επεξεργασίας Εικόνας Orasis (εκδ. 1.1) είναι τώρα ελεύθερα διαθέσιμη για κατέβασμα από τη διεύθυνση:

Μπορείτε τώρα να βελτιώσετε τις φωτογραφίες σας, μόνο με μερικά κλικ.
Το Orasis είναι ένα πειραματικό, βιολογικά εμπνευσμένο, λογισμικό Επεξεργασίας Εικόνας, το οποίο χρησιμοποιεί τα χαρακτηριστικά των κυττάρων ανταγωνισμού κέντρου-περιφέρειας του Ανθρώπινου Οπτικού Συστήματος.
Πολλές φορές η εικόνα που καταγράφει η φωτογραφική μηχανή, και η εικόνα στα μάτια μας, διαφέρουν δραματικά. Ειδικά όταν στη σκηνή που φωτογραφίζουμε υπάρχουν σκιές ή ισχυρές φωτεινές πηγές. Στις περιπτώσεις αυτές, τα μάτια μας διακρίνουν πολύ περισσότερες λεπτομέρειες μέσα στις σκιάσεις και στους έντονους φωτισμούς. Αντίθετα, η εικόνα που μας δίνει η φωτογραφική μηχανή διαθέτει πολύ λιγότερες λεπτομέρειες, έχοντας εκτεταμένες σκοτεινές ή πολύ φωτεινές περιοχές.
Το Orasis προσπαθεί να γεφυρώσει το χάσμα μεταξύ «αυτού που βλέπουμε» και «αυτού που μας δίνει η φωτογραφική μηχανή». Βελτιώνει τις σκιασμένες ή τις πολύ φωτεινές περιοχές, ενώ ταυτόχρονα αφήνει ανέπαφες τις σωστές περιοχές της εικόνας. Το τελικό αποτέλεσμα είναι πολύ πιο κοντά σε αυτό που αντιλαμβάνεται ο άνθρωπος, όταν παρατηρεί τη σκηνή, αποκαλύπτοντας λεπτομέρειες οι οποίες διαφορετικά δε θα ήταν αντιληπτές. Επιπλέον, το Orasis, βελτιώνει την ισορροπία χρωμάτων, τη χαμηλή τοπική αντίθεση και το θόρυβο των φωτογραφιών.