Hessian Optimal Design for image retrieval

Ke Lu, Jidong Zhao and Yue Wu

Available online 19 November 2010.

Abstract

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 http://khup.com/view/18_keyword-image-vectorize/introduction-to-content-based-image-retrieval.html

20 Must-See TED Speeches for Computer Scientists

Article from http://www.mastersincomputerscience.com/20-must-see-ted-speeches-for-computer-scientists/

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 TED.com

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 TED.com

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 TED.com

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 TED.com

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.

5. http://www.ted.com/talks/stephen_wolfram_computing_a_theory_of_everything.html

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.

7. http://www.ted.com/talks/gary_flake_is_pivot_a_turning_point_for_web_exploration.html

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

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 (http://www.isiknowledge.com) 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 (http://www.isiknowledge.com) 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.

@article{Freyne:2010:RSJ:1839676.1839701,
 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 = {http://doi.acm.org/10.1145/1839676.1839701},
 doi = {http://doi.acm.org/10.1145/1839676.1839701},
 acmid = {1839701},
 publisher = {ACM},
 address = {New York, NY, USA},
}

Elsevier Releases Image Search; New SciVerse ScienceDirect Feature

http://www.blogconglomerate.com/headlining-news/elsevier-releases-image-search-new-sciverse-sciencedirect-feature.html

Elsevier a leading publisher of scientific, technical and medical (STM) information, today announced the availability of Image Search, a new SciVerse ScienceDirect (http://www.sciencedirect.com) 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.”

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.