Algorithms for image, video and audio data processing, computer vision
- encoding /decoding - audio & video: wav, mp3, aac, jpeg, mjpeg, h.264, etc;
- augmented reality objects adding (AR);
- image analysis, objects selection, data processing: speed of objects; behavior of objects; changing facilities; comparison with reference; objects tracking; objects dimensioning.
The algorithms' development is a key task in the processing of images, video & audio data, and machine/computer vision. To implement specific functions into devices in an acceptable hardware and software solution, many approximations of the technical system model are required.
Typical tasks: obtaining video and images from specialized devices and equipment; the use of graphic tools for image or video processing, visualization of objects, or data about objects; transfer of algorithms to embedded hardware platforms; development of embedded real-time systems for video processing.
Tasks of improving image quality for visual presentation or analysis; carrying out spatial transformations; detection and measurement of characteristic features of objects in images.
Video processing systems are defined by the architecture of stream processing, for example, in a sequential stream; individual frames are processed one by one or several per period. Such processing is essential for systems using video streams, or if the size of video files is too large to load it for processing entirely.
Video capture from various devices - IR sensors, thermal images, web cameras, and industrial video modules, etc. Solutions for preliminary and subsequent processing - noise interference, low contrast, defocusing of the optics. Carrying out video analysis using various methods, for example, edge selection, blob analysis, pattern matching, optical flow and angles detection.
Development of solutions for common video processing tasks, such as stabilization, object detection, object tracking, object dimensioning, comparison of contours or other parameters with reference ones.
Spatial calculations - analysis of positional data and visualization of spatial objects; import and export of vector and raster data from / to various file formats; transformation of data by selecting image areas, trimming, overlapping, adjusting spatial resolution; visualization of data combining information about objects and base layers from multiple sources; measurement of objects, their dimensions and distances between objects and objects, calculation of distances, heights, areas, three-dimensional transformation of coordinates and construction of projections of objects.