Ievgen G. Profile | Freelancer

Como Isso Funciona

Busque Trabalhos

Publicar um Projeto

5,0

(93 avaliações)

Foto de capa do perfil

Você está agora seguindo

Erro seguindo usuário.

Esse usuários não permite que o sigam.

Você já está seguindo esse usuário.

Seu plano permite apenas 0 seguidas. Aprimore-o aqui.

Deixou de seguir com sucesso

Ocorreu um erro ao deixar de seguir o usuário.

Você recomendou com sucesso

Ocorreu um erro ao recomendar o usuário.

Algo deu errado. Por favor, atualize a página e tente novamente.

E-mail verificado com sucesso

$100 USD / hora

Bandeira do(a) UKRAINE

kharkiv,

$100 USD / hora

Atualmente, está 10:34 AM aqui

Ingressou em agosto 16, 2011

25 Recomendações

Ievgen G.

@Zhats

5,0

(93 avaliações)

9,5

9,5

93%

93%

$100 USD / hora

Bandeira do(a) UKRAINE

kharkiv,

$100 USD / hora

99%

Trabalhos Concluídos

100%

Dentro do Orçamento

99%

No Prazo

30%

Taxa de Recontratação

PhD, computer vision, AR/VR, SLAM

I’m a tech entrepreneur with passion for computer vision and machine learning, PhD degree in image processing and 10+ years of R&D experience. I have successfully built a strong team of experts. Most of them have grown under my direct supervision at It-Jim. Currently I’m managing multiple R&D and software engineering groups. We are delivering consulting services in computer vision, pattern recognition, machine learning, signal and image processing, augmented and mixed reality. As a researcher and leader, i'm always striving to solve challenging problems and deliver the highest quality of services. This is the main spirit of It-Jim company. You can find more details here [login to view URL]

Ver todos Matlab and Mathematica Engineers em Ukraine

Contate Ievgen G. sobre seu trabalho

Faça login para discutir quaisquer detalhes via chat.

Portfólio

Automatic document analysis is a key part of the overall document recognition process. A common scenario is when the user takes a picture by mobile phone or tablet and the goal is to automatically parse and recognize content from the captured document. In this project, we have built the page unwrapping engine. The developed algorithm contains several key components: Preprocessor, Feature Extractor, Geometric Model Estimator and Refiner. The preprocessor performs image filtering, roughly locates the document boundaries and extracts vertical and horizontal spans (lines and text regions).Feature Extractor module performs parsing of document content. For robustness, we have fused several types of features including corners and edges from images preprocessing in different ways. The output nonlinear grid of points is used as an input for the next system component: Geometric Model Estimator. Finally, we receive the corrected document.

Document Analysis and Recognition

In this project, our goal was to develop an engine for automatic counting of the various symbols on scanned building plans. Our custom algorithm was able to handle hundreds of different symbols, including complex ones consisting of multiple small sub-patterns and text. It is robust to rotation, scaling and partial symbol occlusions. The developed engine is based on a purely classic computer vision algorithm, works very fast and does not require GPU, which makes it very cheap to use in the cloud. In addition, a comparative test with market competitors indicates an advantage from both accuracy and performance points of view.
KEY POINTS:
1. Automatic recognition of symbols with different complexity
2. Optimized search using multi-threading
3. Custom decomposition algorithm
4. Rotation and scale-invariant recognition engine
5. Successfully integrated into an engineering platform

Symbol Recognition

We have built a large-scale system for real-time sports analytics. The main goal of the project was to replace the human operator and automatically find the most interesting area on the hockey/football/ basketball pitch.
The region-of-interest (ROI) control algorithm required pitch mask segmentation for avoiding false detections of the player blobs, panorama stitching from multiple cameras in real time, adaptive filtering and thresholding to detect the player candidates and custom tracker for robust keeping the traces of players. A specific clustering procedure was developed in order to build a proper logic for selection and smooth control of the ROI.
The key advantage of the created tracker is extremely high performance. It works at 120 FPS and multiple system instances can be simultaneously installed in the cloud. This makes the solution quite cheap and flexible, ROI tracking and streaming of the outputs are performed for many clients all over the globe on a real-time basis.

Hockey Tracker for Sports Applications

Visual image search and content-based image retrieval (CBIR) is a very hot topic in computer vision and machine learning communities. It is a core of many systems in retail, entertainment, biometrics, robotics, and other domains. In our project, we have successfully developed a mobile image retrieval system for augmented reality applications. For this purpose, we have optimized and customized the bag-of-words (BoW) algorithm using a properly constructed hierarchical tree for the features vocabulary. As a result, local features from the test images were matched extremely fast. Our CBIR system worked on a mobile device without an internet connection and returned the visual search results within only 2 milliseconds.

KEY POINTS:
1. Instant retrieval offline on mobile device
2. 2 ms time for 1000 images database
3. Works on iOS and Android
4. Integrated into custom mobile AR SDK

Mobile Image Retrieval

Drones are widely used for various tasks including visual inspection, facades and buildings reconstruction, surveillance and tracking, object detection and recognition, etc. In this project, we have built a web-tool for visual inspection of buildings. The core technology was a customized 3D reconstruction pipeline working with both image sequences or videos from the drone camera. We have implemented structure-from-motion (SfM), point cloud densification, meshing and texturing modules. The resulting pipeline was integrated into the web-service, where the user was able to load the data for reconstruction, analyze the resulting 3d model, view the fine details and save the results for further steps.
KEY POINTS:
1. 3D point cloud reconstruction engine is implemented
2. Successfully integrated into web-service
3. Full pipeline including SfM, meshing and texturing is created

3D Inspection by Drones

Marker-based augmented reality (AR) is extremely popular nowadays. However, full user immersion is possible only in the case of robust real-time computer vision solutions working on the mobile device. We have developed a custom hybrid tracking system based on local feature tracking and template-based matching. The tracking engine tracks the homography changes using an optical flow algorithm and then refines the residual warp using the optimized template matcher. Developed tracker is robust to extremal angles, drift-free and very accurate. Algorithms work at 60 FPS on Android and iOS.
KEY POINTS:
Custom tracker is resistant to extremal angles and drift
60 FPS on mobile device (iOS+Android)
Fusion of feature-based and template-based trackers provides high robustness
Successfully used in custom mobile AR SDK

Mobile Augmented Reality SDK

Automatic document analysis is a key part of the overall document recognition process. A common scenario is when the user takes a picture by mobile phone or tablet and the goal is to automatically parse and recognize content from the captured document. In this project, we have built the page unwrapping engine. The developed algorithm contains several key components: Preprocessor, Feature Extractor, Geometric Model Estimator and Refiner. The preprocessor performs image filtering, roughly locates the document boundaries and extracts vertical and horizontal spans (lines and text regions).Feature Extractor module performs parsing of document content. For robustness, we have fused several types of features including corners and edges from images preprocessing in different ways. The output nonlinear grid of points is used as an input for the next system component: Geometric Model Estimator. Finally, we receive the corrected document.

Document Analysis and Recognition

In this project, our goal was to develop an engine for automatic counting of the various symbols on scanned building plans. Our custom algorithm was able to handle hundreds of different symbols, including complex ones consisting of multiple small sub-patterns and text. It is robust to rotation, scaling and partial symbol occlusions. The developed engine is based on a purely classic computer vision algorithm, works very fast and does not require GPU, which makes it very cheap to use in the cloud. In addition, a comparative test with market competitors indicates an advantage from both accuracy and performance points of view.
KEY POINTS:
1. Automatic recognition of symbols with different complexity
2. Optimized search using multi-threading
3. Custom decomposition algorithm
4. Rotation and scale-invariant recognition engine
5. Successfully integrated into an engineering platform

Symbol Recognition

We have built a large-scale system for real-time sports analytics. The main goal of the project was to replace the human operator and automatically find the most interesting area on the hockey/football/ basketball pitch.
The region-of-interest (ROI) control algorithm required pitch mask segmentation for avoiding false detections of the player blobs, panorama stitching from multiple cameras in real time, adaptive filtering and thresholding to detect the player candidates and custom tracker for robust keeping the traces of players. A specific clustering procedure was developed in order to build a proper logic for selection and smooth control of the ROI.
The key advantage of the created tracker is extremely high performance. It works at 120 FPS and multiple system instances can be simultaneously installed in the cloud. This makes the solution quite cheap and flexible, ROI tracking and streaming of the outputs are performed for many clients all over the globe on a real-time basis.

Hockey Tracker for Sports Applications

Visual image search and content-based image retrieval (CBIR) is a very hot topic in computer vision and machine learning communities. It is a core of many systems in retail, entertainment, biometrics, robotics, and other domains. In our project, we have successfully developed a mobile image retrieval system for augmented reality applications. For this purpose, we have optimized and customized the bag-of-words (BoW) algorithm using a properly constructed hierarchical tree for the features vocabulary. As a result, local features from the test images were matched extremely fast. Our CBIR system worked on a mobile device without an internet connection and returned the visual search results within only 2 milliseconds.

KEY POINTS:
1. Instant retrieval offline on mobile device
2. 2 ms time for 1000 images database
3. Works on iOS and Android
4. Integrated into custom mobile AR SDK

Mobile Image Retrieval

Drones are widely used for various tasks including visual inspection, facades and buildings reconstruction, surveillance and tracking, object detection and recognition, etc. In this project, we have built a web-tool for visual inspection of buildings. The core technology was a customized 3D reconstruction pipeline working with both image sequences or videos from the drone camera. We have implemented structure-from-motion (SfM), point cloud densification, meshing and texturing modules. The resulting pipeline was integrated into the web-service, where the user was able to load the data for reconstruction, analyze the resulting 3d model, view the fine details and save the results for further steps.
KEY POINTS:
1. 3D point cloud reconstruction engine is implemented
2. Successfully integrated into web-service
3. Full pipeline including SfM, meshing and texturing is created

3D Inspection by Drones

Marker-based augmented reality (AR) is extremely popular nowadays. However, full user immersion is possible only in the case of robust real-time computer vision solutions working on the mobile device. We have developed a custom hybrid tracking system based on local feature tracking and template-based matching. The tracking engine tracks the homography changes using an optical flow algorithm and then refines the residual warp using the optimized template matcher. Developed tracker is robust to extremal angles, drift-free and very accurate. Algorithms work at 60 FPS on Android and iOS.
KEY POINTS:
Custom tracker is resistant to extremal angles and drift
60 FPS on mobile device (iOS+Android)
Fusion of feature-based and template-based trackers provides high robustness
Successfully used in custom mobile AR SDK

Mobile Augmented Reality SDK

Avaliações

Mudanças salvas

Mostrando 1 - 5 de 50+ avaliações

Filtrar avaliações por:

5,0

$3.000,00 USD

Project for Zhats -- 2

Perfect!

Matlab and Mathematica

Financial Research

+1 mais

Björn W.

@Media4You

•

há 6 anos

5,0

$473.520,00 USD

Project for Zhats -- 3

Perfect!

Matlab and Mathematica

C++ Programming

+1 mais

Björn W.

@Media4You

•

há 6 anos

5,0

$4.104,00 USD

Image Processing XC

He's very intelligent and delivered large scale project at very high quality. He's also very flexible and very responsive. Moreover, he's patient to the extent of walking through every bit of work.

Matlab and Mathematica

P

Omar Z.

@Projectorist

•

há 7 anos

5,0

$1.700,00 USD

Project for Zhats -- 3

Great to work with. Extremely intelligent and dependable. Will continue to work with Zhats on future projects.

Matlab and Mathematica

C++ Programming

+1 mais

H

Hannibal B.

@hannibalbaldwin

•

há 8 anos

5,0

$2.000,00 USD

Project for Zhats -- 2

Zhats is a exceptional freelancer He is very motivated and is able to achieve beyond your expectation We are returner customers and very happy to continue our relation with him

Matlab and Mathematica

C++ Programming

+1 mais

C

Closed User

@ClienDrew

•

há 8 anos

Educação

PhD in mathematics and physics

Ukraine

2010 - 2013

(3 anos)

M.Sc.

Harkivs'kij Nacional'nij Universitet im. V.N. Karazina,

Ukraine

2005 - 2010

(5 anos)

Qualificações

Warsaw University of Technology

2015

The award is given for the best presentation of scientific results. Two papers of mine have won simultaneously: 1. Efficient Data Focusing and Trajectory Reconstruction in Airborne SAR systems 2. Robust Numper Plate Detector based on Stroke Width Transform and Neural Network

YSC-2014

2014

The award for the best paper in the young scientists conference

MMET

2012

The 1st prize for the best paper on the international conference "Mathematical Methods in Electrodynamics".

Publicações

Multi-Look Stripmap SAR Processing Algorithm with Built-In Correction of Geometric Distortions

Proceedings of the European Conference on Synthetic Aperture Radar (EuSAR'2010)

An efficient time-domain algorithm for synthetic aperture radar (SAR) is proposed. The advantage of given method is an improved multi-look processing scheme with possibility to obtain the radar image not in ''radar coordinates', but directly on the ground plane.

Range-Doppler Algorithm with Extended Number of Looks

Proceedings of the 2011 Microwaves, Radar and Remote Sensing Symposium

The Range-Doppler Algorithm (RDA) uses the Discrete Fourier Transform (DFT) for the efficient calculation of the convolution in the frequency domain. In the proposed method we extend the processed Doppler bandwidth in order to control real antenna beam deviation and nonuniform scene illumination.

Estimation of Phase Errors in SAR Data by Local-Quadratic Map-Drift Autofocus

Proceedings of 13th International Radar Symposium

We have developed an efficient autofocusing approach, which allows to retrieve an arbitrary phase error from radar data. This algorithm gives a possibility to use quite cheap GPS receiver for the aircraft trajectory measurement. Proposed LQMDA method is used for high-resolution imaging up to 0.5 meter resolution.

Contate Ievgen G. sobre seu trabalho

Faça login para discutir quaisquer detalhes via chat.

Verificações

Freelancer Preferencial

Identidade Verificada

Pagamento Verificado

Telefone Verificado

E-mail Verificado

Conectado ao Facebook

Certificações

US English 1

80%

Principais Habilidades

Matlab and Mathematica

75

38

25

C++ Programming

24

2

Busque Freelancers Parecidos

Matlab and Mathematica Engineers in Ukraine

Matlab and Mathematica Engineers

Algorithm Experts

Busque Mostruários Parecidos

Matlab and Mathematica

C++ Programming

Usuário Anterior

Usuário Seguinte

Convite enviado com sucesso!

Obrigado! Te enviamos um link por e-mail para que você possa reivindicar seu crédito gratuito.

Algo deu errado ao enviar seu e-mail. Por favor, tente novamente.

Carregando pré-visualização

Permissão concedida para Geolocalização.

Sua sessão expirou e você foi desconectado. Por favor, faça login novamente.