Project System

V ladislavs DOVGALECS

www.dovgalecs.com

** *** ******* ** **********

***** *****

FRANCE

R ésumé

Personal details P rofessional profile

Date of Bi rth: 27.06.1984 Research and development of

Citizenship: L atvian p ractical applications in multimedia

E-mail: using Pattern Recognition and Machine

**********.*********@*****.*** Learning. Implemented systems and

Phone: +33 6 43 18 10 55 i ntelligent modules for visual

i nformation search, recognition and

classification by content.

Cu r rent situation

Developing algorithms for advanced image-based retrieval system for rapid

search in digitized historical documents. The DocExplore project, L IT IS,

U niversity of Rouen, France.

Education

• PhD – U niversity of Bordeaux, France (Sciences of Physics and

E ngineering)

• Master – U niversity College of Ventspils, Latvia (Math. Fundamentals of

CS)

• Bachelor – U niversity College of Venstpils, Latvia (Computer Sciences)

Summa ry of work experience

2011 – present

Postdoctoral fellow in University of Rouen, France. Contributing to the

DocExplore project. Developing an advanced image-based search and

retrieval subsystem as a part of the innovative DocExplore project software

suite.

Keywords: I nformation Ret r ieval, M achine Lea rning and P atte rn

R ecognition

2008 – 2011

PhD candidate in University of Bordeaux, France. Contributed to the

I M MED p roject. Designed and developed a system for indoors localization

using pure i mage-based techniques. The system provided essential clues for

person activity recognition system.

Keywords: A lgorithms, D ata M ining, C lassification, I mage-based

L ocalization

2007 – 2008 (in parallel to studies)

Computer and Network administrator (UN IX systems), IT systems

management. Designed and implemented two large network sites – (1)

hardware solutions and (2) network services. Developed storage and web

solutions for local and public needs.

Keywords: T CP/ IP, L A N, WAN, U N IX/Linux, S amba, A pache, M ySQL

2006 – 2008 (in parallel to studies)

Junior specialist in Engineering and Research Center in

Telecommunications in Ventspils, Latvia. Contributed to various

telecommunication projects (IP Television, video surveillance). Conducted a

search for necessary equipment and created detailed specifications for

l arge-batch order.

Keywords: I P T V quality assessment and v ideo surveillance solutions

Skills

Machine Lea rning & Data mining

• Mastering inference and prediction algorithms such linear & nonlinear

methods for classification (kernel machines, decision t rees, Bayes

classifiers) and regression;

• Mul tiple information source fusion

• Unsupervised learning in high dimensionality data (cluster analysis,

d imensionality reduction, prior information integration)

• Machine learning for low-power embedded devices (Extreme Learning

M achines)

• Understanding in algorithm complexity assessment (Big-Oh notation,

memory usage optimization)

P rogramming & Development

MATLAB - strong skills in prototype creation and experimentation with

data

C++/JAVA – average skills in scientific computation (e.g. parallel

computing), Object-Oriented programming (STL, Boost), OpenCV,

I n tel IPP & TBB

Python, Bash - good skills

Collaboration tools – experienced with SVN, Git, CVS

IDE – NetBeans, Eclipse

Computer OS

Advanced user of UN IX/Linux/Mac OS X systems

Experienced with network technologies and services

Embedded system design and programming (advanced use of Raspberry

P i)

Electronics

In termediate / hobbyist level in analog and digital circuits

Mul tiple sensor data processing (accelerometers, gyroscopes, vision)

Assembler & C programming for embedded devices (microcontrollers,

ARM)

Languages

Latvian / Ukrainian – Native speaker

English / Russian – Fluent speaker

French - Good level

D etails on wor k experience

2011 – cur rent

Engaged as Postdoctoral research fellow at University of Rouen, I am

participating in the DOCEXPLORE project aiming to preserve and render public

t he r ich historical heritage in form of handwri t ten medieval documents.

My research concerns design and development of advanced algorithms and

tools for image-based search in historical handwri t ten documents. I have

i mplemented novel algorithms and tools for arbitrary word or graphical pattern

search in handwrit ten documents.

During one year of research and development, the research results were

validated and published in scientific articles as well as implemented in several

p rototypes.

See h ttp://www.docexplore.eu for more details.

References:

Laurent HEUTTE (*******.******@****-*****.** )

Pierrick TRANOUEZ (********.********@****-*****.** )

Stéphance N ICOLAS (********.*******@****-*****.** )

2008 – 2011

PhD Thesis: “Indoor Location Estimation using a Wearable Camera with Application to

the Monitoring of Persons at Home”.

Video information is a rich source of information and can be used in a large variety of

practical application – for instance patient monitoring in medical context. Localization and

activity information are the strong cues for treating physician attempting to detect the

Alzheimer disease.

During three years of research, I developed novel algorithms for HD quality video

indexing by topological location using purely vision-based pattern recognition approaches.

The work resulted in several scientific publications as well as a working prototype for medical

personnel involved in the validation process. Contributed in light-weight prototype

development and evaluation.

See http://immed.labri.fr for more details.

References:

Yannick BERTHOUMIEU (*******.***********@***-********.**)

Rémi MEGRET (****.******@***-********.**)

2006 – 2008

Master Thesis: “Mammogram segmentation and characterization using iterative wavelets

algorithms”, Nancy University, INPL, France.

ERASMUS exchange student in Nancy (INPL), France. Grade : 16 of 20

Breast cancer is a serious threat to women health in many developed countries with

higher expected age. Routine scanning proved to be an efficient measure for prevention and

early disease detection. Nevertheless, mammogram images are difficult to analyze rapidly,

hence the need for automated image processing solutions.

During 5 month internship I was introduced to practical research and proposed a series

of relatively simple image filters for micro-calcification detection in mammogram images

using wavelet techniques.

References:

Radu Ranta (****.*****@*****.****-*****.**)

Christian Daul (*********.****@*****.****-*****.**)

2003 – 2006

Bachelor thesis: “ M ic rocontrollers in engine control and d rive gear system

of I rbene RT-32 radiotelescope ”.

Grade: 9 of 10

Radioastronomic observations provide an insight in processes happening at universe

scale. To obtain meaningful results, precise measurement tools are necessary.

I studied the motor control system of the former soviet military radiotelescope RT-32

(Latvia, Irbene) and proposed a conceptual design based on microcontrollers. The

radiotelescope is currently participating in various European Very Long Base Interferometry

projects.

See http://virac.venta.lv/en/ for more details.

Contact this candidate