In our research laboratory at Poznan University of Technology, we process data for a wide range of perception applications using a variety of vision sensors and hardware platforms, employing computer vision and machine learning algorithms and techniques.
Intelligent computer vision and signal processing
We work at the intersection of state of the art methods of image processing and machine learning to extract meaningful information from images, videos and signals. Our expertise covers a wide range of tasks including (among others) image filtering and enhancement, segmentation, navigation, object detection and tracking for applications in robotics, monitoring, remote sensing, agriculture and healthcare.
Perception beyond visible spectrum
Aside from regular camera images, we work with a wide variety of imaging sensors using different modalities including depth and range data, thermal imaging data, multispectral and hyperspectral images to to find the best, result-driven solution for the target application.
Edge AI and embedded systems
Part of our research work is also the deployment of algorithms on edge AI embedded devices, ranging from Linux-based single board computers and programmable hardware system-on-chip devices to tiny microcontrollers. The range of expertise covers adjustment of algorithms for the target architecture deployment including low-level optimizations, quantization, mixed precision processing and the use of efficient deployment runtimes.