CBIA — Centre for Biomedical Image Analysis

The Centre for Biomedical Image Analysis (CBIA) is a well-established, interdisciplinary unit that attends to research, education as well as service activities related to automated image analysis applied mostly in biology and medicine. It primarily focuses on the development and benchmarking of algorithms for the analysis and synthesis of biomedical image data as well as on the employment of computers in the optimization and automation of the biomedical image acquisition process.

CBIA carries out its research in close collaboration with experts in biology and medicine. With prevention, diagnostics and therapy in oncology being our main application areas, we aim at describing the spatiotemporal behavior of cancer cells and tumor tissues under various conditions. To this end, we develop reliable and ideally automatic detection, segmentation, classification, tracking and quantification algorithms, mostly by combining machine learning and traditional image analysis approaches. For benchmarking purposes, we attend to the research of data and annotation quality, data augmentation techniques as well as metrics and ranking strategies for measuring algorithm performance.

CBIA coordinates the national Centre of AI in Oncology and closely collaborates with other relevant groups mostly across Europe within pan-European research infrastructure projects like ESFRI Euro-BioImaging (on national level Czech-BioImaging) or COST NEUBIAS. CBIA also co-organizes the Cell Tracking Challenge that has established an internationally widely recognized international benchmark for objective comparison of cell segmentation and tracking algorithms. Over the 10-year existence of this benchmark, more than 90 approaches have been evaluated and detailed analyses of their behavior have been published in Nature Methods (2017 and 2023).