Ph.D. Dissertation Proposal

A Rapidly Deployable Image
Classification System Using Multi-Views

Adrian Rosebrock

11:00am Friday, 10 May, ITE 325, UMBC

Constructing an image classification system using strong, local invariant descriptors is time consuming and tedious, requiring many experimentations and parameter tuning to obtain an adequately performing model. Furthermore, training a system in a given domain and then migrating the model to a separate domain will likely yield poor performance. As computer vision systems become more prevalent in the academic, government, and private sectors, it is paramount that a framework to more easily construct these classification systems be created. In this work we present a rapidly deployable image classification system using multi-views, where each view consists of a set of weak global features. These weak global descriptors are computationally simple to extract, intuitive to understand, and require substantially less parameter tuning than their local invariant counterparts. We demonstrate that by combining weak features with ensemble methods we are able to outperform the current state-of-the-art methods or achieve comparable accuracy. Finally, we provide a theoretical justification for our ensemble framework that can be used to construct rapidly deployable image classification systems called "Ecosembles".

Committee: Dr. Tim Oates (chair), Dr. Jesus Caban, Dr. Tim Finin, Dr. Charles Nicholas

Classes each Wednesday evening starting July 8.

This five week workshop prepares participants to take the computer-adaptive Graduate Records Examination. The purpose of the course is to improve students' performance on the GRE by:

• Providing students with successful test management techniques
• Familiarizing students with question types specific to the GRE
• Teaching the logical, rhetorical, and mathematical problem-solving techniques relevant to the GRE
• Teaching essential grammatical and mathematical concepts and techniques in application to real test problems
• Identifying the specific problems posed by the GRE, and analyzing common test-taking errors
• Improving performance and reducing test anxiety through practice testing

This workshop is taught by Dr. Anthony Poplawski, who has helped thousands of students perform better on their GRE examinations. While this workshop provides participants with very valuable information, techniques and strategies for maximizing their potential on the GRE, please note that it does not replace studying. In addition to taking this workshop we recommend that students dedicate significant time to studying for the examination in order to be truly prepared.

Please visit the Graduate Record Examination web site at www.gre.org to view information on the GRE exam.

TO REGISTER AND LEARN MORE ABOUT THE COURSE, PLEASE VISIT:

http://www.umbc.edu/testprep/gre.html

Take advantage of early bird discounts by registering before June 17, 2013!

Medifast is looking to hire UMBC IS, CompSci, BTA and EE majors for summer positions in Owings Mills, MD.  Get involved with this fantastic company!  Position descriptions are below.  To apply, make sure to click the link on each position title to get directed to the Medifast website.  Get your credentials in to them today!

IT Business Analyst Intern: - IS, CompSci, BTA & EE Majors

Job Description:

As part of the I.T. Project Team, the Business Analyst Intern will participate in project meetings and discussion and based on the teams input, and with direction, take ownership of creating the Business and Functional Documentation based on a best practice template.

The Business Analyst Intern will be part of the project team and support the team through the phases of the project by gathering information, seeking answers to question, creating process flow diagrams and following up on open items.

Knowledge and Skills:

The candidate should have experience with or exposure to the following technologies and skills through hands on experience and/or relevant course work:

• Excel
• Word
•  Written and verbal communication skills
• Good comprehension skills
• Ability to adapt to changing priorities

Education:

Viable candidates must be currently enrolled in a degree college/university program working toward a degree in Computer Science, Electrical Engineering, Computer Engineering or other Information Technology related discipline. Must be able to work 15-20 hours per week during the school year and 40 hours during breaks.

Reporting Analyst Intern - FIEC and ECON Majors

Medifast is looking to hire an intern to join our Client Solution Center Operations & Technology Support team for Summer 2013. This person will take on a Reporting Analyst role to conduct data analysis on performance metrics for our call center agents.

• Update reports, spreadsheets and databases daily
• Modify model based on learned information to allow for more precise volume and staffing projections
• Identify objective and reportable metrics for balanced scorecard
• Tie the metrics to areas of delivery that may be impacted and contributing to negative results to derive actionable items for resolution
• Identify and extrapolate information from Business Intelligence data sources where appropriate

Skills Needed:

•  Advanced level Experience using Excel including data links; VLookup, Pivot Tables, advanced formula, macros, etc.
• Experience conducting interviews to determine existing data sources to create output to meet business needs
• Ability to assess current reporting methods and identify ways to produce quality and accurate reports with greater efficiency and broader output options
• Must be able to create and manage Access Database
• Ability to import source data and create reports
• Adept at queries and report writing
• Ability to train staff on data management techniques
• Proficiency using Word and PowerPoint

Internships have become a must for college students; without the work experience they provide, the post-graduation job search is significantly more difficult. But it’s not enough to simply get an internship. You also need to impress the employer during your time there.

Here are 10 ways to get the most out of an internship:

1. Know what to expect. Generally, the idea behind an internship is to give you some basic exposure to day-to-day work in your field. In most cases, you will not be doing glamorous, substantive work; you’ll be making other people’s lives easier. This means you may get stuck photocopying, filing, arranging meetings, and completing other tasks that may strike you as drudgery. In exchange, you get exposure to the field and work experience to put on your resume.

2. Gain trust early on. If you excel at the boring tasks and do them cheerfully, you may be given more interesting work. Now, you may wonder what being good at photocopying has to do with your ability to do, say, independent research. Here’s the connection: When you come in as an intern, you haven’t proven yourself in the work world. But if you do a great job on the boring work, you’ll show that you pay attention to detail, follow instructions, and care about quality. Keep up that track record, and eventually someone may let you try something more interesting. But do a bad job on the basic stuff, and no one will trust you with anything more advanced. So it’s important to go into the job determined to do every task well, no matter how menial.

3. Pay attention to the office culture. Observe how others in the office act, and mirror that. For instance, if employees modulate their voices when others are on the phone, modulate yours. If they’re compulsively on-time for meetings, you should be compulsively on-time, too. Lots of little things like this will help you appear professional. And while they may sound small, they’re likely to help you stand out compared to other interns.

4. Focus. Don’t use social networking sites (unless it’s part of your job) or text with friends throughout the workday. You may be confident that it doesn’t distract you or affect your work, but experienced managers have watched enough people to be confident that it does.

5. Take your work seriously. In school, if you made a mistake on a test or paper, it only affected you. In many jobs, mistakes are much more serious. If you do make a mistake, make sure you handle it correctly.

6. Ask for feedback. Every so often, ask your boss how you’re doing. What could you be doing differently? Make it easy for her to give you input that will help you grow.

7. Learn from your co-workers. Ask them about their own careers. How did they get into the field? What do they like about it? What do they find challenging? What advice do they have for you? Most people love to talk about themselves and will be flattered that you’re asking about their experiences. Best of all, it’s likely to make them want to help you.

8. Dress appropriately. There’s no “intern exception” in the dress code, and yet I’ve seen interns come to work wearing flip-flops, ultra-low-rise jeans, visible bra straps, and worse. If you look like you're dressing for a class rather than a job, you’re signaling that you don't take your job seriously.

9. Ask for advice. Talk to people about your career plans. Tell them you'd love any advice they have, either now or in the future. Your co-workers can be helpful to you by telling you about job leads, recommending you for a job, and helping you consider various career choices. Though most people are happy to offer this kind of help, they might not offer it if you don't explicitly ask.

10. Say “Thank you.” Talk to your manager about what you’re getting out of your internship, and thank her for giving you the opportunity to work there. We all love hearing the occasional expression of appreciation, so don't be shy about offering it.

Name:  Babak Vint

Internship, Co-op or Research Site:  Beck Radiological Innovations

Location of the Organization (City, State):  Halethorpe, MD

Title of Your Position:  Biomedical Engineering Assistant

Major(s)/Minor(s):  Mechanical Engineering/Computer Science

The thing I have enjoyed most about this position is the sense of having contributed to medical technology.  It would be simply ideal if medical technology was cheap and accessible, because it would drive down the cost of healthcare.

CSEE graduate student Asmita Korde will present a paper on her research with Professor Tinoosh Mohsenin tomorrow at the the SPIE Defense Security and Sensing Conference in the Baltimore Convention Center. Asmita was a UMBC CWIT Scholar and received a BS degree in Computer Engineering in 2011. She is now finishing her MS degree at UMBC in the Electrical Engineering program.

Her paper, Detection Performance of Radar Compressive Sensing in Noisy Environments, describes research done in collaboration with her mentor, Tinoosh Mohsenin, and Damon Bradley of the NASA Goddard Space Flight Center.

Here is the abstract.

In this paper, radar detection via compressive sensing is explored. Compressive sensing is a new theory of sampling which allows the reconstruction of a sparse signal by sampling at a much lower rate than the Nyquist rate. By using this technique in radar, the use of matched filter can be eliminated and high rate sampling can be replaced with low rate sampling. In this paper, compressive sensing is analyzed by applying varying factors such as noise and different measurement matrices. Different reconstruction algorithms are compared by generating ROC curves to determine their detection performance. We conduct simulations for a 64-length signal with 3 targets to determine the effectiveness of each algorithm in varying SNR. We also propose a simplified version of Orthogonal Matching Pursuit (OMP). Through numerous simulations, we _nd that a simplified version of Orthogonal Matching Pursuit (OMP), can give better results than the original OMP in noisy environments when sparsity is highly over estimated, but does not work as well for low noise environments.

About the program:

The Department of Biological Sciences and the Department of Mathematics and Statistics announces the Undergraduate Mathematics Biology Training Program (ubm@umbc.edu) supported by the National Science Foundation. At the core of this program is the belief that modern scientific problems demand multidisciplinary approaches, and that cross-disciplinary training will accelerate discovery and prepare students to be scientific leaders in the 21st century. Our innovative program will provide two years of research training for undergraduates working on projects at the interface of biology, mathematics and statistics.

For student candidates:

The goal of this program is to prepare undergraduate students for graduate study and careers in biomathematics, biostatistics and fields that require a strong background in computational approaches to biological problems.

Apply NOW!