Sviatoslav Ponomarev Computer Vision Engineer

Professional Experience

Huawei department, Ventra Employment

Senior Engineer

February 2020 - Present

  • Developed fast and robust algorithms for detection and removal of flickering caused by LED captured under short exposure time [Python, C++, OpenCV]
  • Single camera, stereo camera, camera + IMU systems calibration, calculation of optical systems parameters for generation of more realistic synthetic images [C++, OpenCV, Matlab]
  • Managed a team of 2 people to prepare photorealistic synthetic scenes and generate training set of images with supplemental data (optical flows, occlusion maps)
  • Improved detail/color transfer neural networks accuracy by developing procedural pipeline for photorealistic training images generation for system of cameras with different optical parameters [Python, PyTorch, Blender]
  • Improved robustness of video stabilization by developing global motion estimation algorithms modifications [Python, OpenCV]

Huawei Technologies Co. Ltd.

Individual Businessman Contractor

August 2018 - January 2020

  • Improved visual quality of images captured in low light conditions by developing RGB+NIR images merging/fusion algorithm [Python, OpenCV]
  • Development of color restoration algorithms in case of low light capturing conditions [C++, OpenCV]

Pavlov Institute of Physiology, Russian Academy of Science

Junior Researcher

July 2017 - December 2020

  • Achieved 79% accuracy for automatic classification of visual stimuli based on electroencephalogram analysis and SVM training with wavelet-based features [Matlab]
  • Capturing and creation of an openly available ORDSLAM database to compare the performance of RGB-D SLAM algorithms outdoors [C++, OpenCV]

My Sales Team

Researcher (freelance)

September 2015 - February 2016

  • Developed a goods counting system on store shelves with average recognition accuracy of 85% per goods category [C++, OpenCV]
  • Managed a team of 2 people to conduct a study of object classification algorithms for goods images in stores

Vavilov State Optical Institute

Software Engineer

September 2013 - July 2017

  • Engaged in the development of system for determining the orientation of the observed object in 3D space using stereo vision methods [C++, OpenCV, OpenGL, PCL]
  • GUI development for monitoring program of malfunction of electrical equipment and writing the functional specification [C++, Qt]
  • Development of framework for the performance estimation of object recognition algorithms which speeded up neural network training process [C++, Python, Caffe]
  • Development of mobile phone video recording effects simulation system (blur, rolling shutter effects) [C++, OpenCV]

ITMO University

Engineer

October 2012 - June 2013

  • Improved biometric identification accuracy using 2D and 3D face images by developing modification of stereo vision algorithms [C++, OpenCV]


Education

ITMO University

M.Sc. Photonics and Optical Information Technology

2013 - 2015

Faculty: Photonics and Optical Information Technology
Grade: 4.95
Thesis: Development of the 3d structural image matching algorithm

ITMO University

B.Sc. Photonics and Optical Information Technology

2009 - 2013

Faculty: Photonics and Optical Information Technology
Grade: 4.5
Thesis: Dense stereo vision algorithms for images of three-dimensional scene