Image processing is necessary for a wide array of technological and scientific applications. Optical or electron microscopy, static or dynamic x-ray and neutron imaging, measuring system dynamics with frame sequences acquired via cameras – all of these methods require noise filtering, artefact removal, segmentation analysis, object tracking, parameter measurements etc. Sometimes image restoration is needed especially for damaged and blurred images. In scientific measurements, one often needs to work with images with a low signal-to-noise intensity ratio, which means that specialized methods (often ad hoc) are necessary for non-destructive image filtering.
We offer our services in the field of scientific image processing and image restoration, including processing and analysing data derived from images. We use either existing tools or develop a framework tailored to a certain application.
For image processing, we use proprietary packages Wolfram Mathematica and Matlab (via Matlink), and open-source tools such as ImageJ and OpenCV, Python, etc. We use the following methods for image processing: noise reduction with brightness anisotropic diffusion and curvature flow, BM3D, (non-)local averaging, Fourier and wavelet filters; n-Otsu, Chan-Vese, local adaptive segmentation; texture restoration using synthesis and evolution equation methods, as well as our in-house algorithms such as multiscale recursive interrogation filtering and colour tone map masking.
Cooperation contract with Helmholtz-Zentrum Dresden-Rossendorf (Germany): the development of image processing methods and algorithms for two-phase MHD flow analysis.
- Cooperation agreement with SIA Vizulo (Latvia): Development of ray tracing models for industrial luminaires and optimization of spatial illumination patterns.
- ERDF project: Development of numerical modelling approaches to study complex multiphysical interactions in electromagnetic liquid metal technologies.