Researchers at Warsaw University of Technology observe living cells
Researchers at Warsaw University of Technology observe living cells
Cell observation forms the foundation of biological and medical research. While microscopes have been used for centuries, observing cells proves surprisingly challenging - they are essentially transparent and invisible to conventional microscopy. Staining cells makes them visible but potentially alters their natural state, leaving researchers uncertain whether observed behaviours reflect genuine cellular activity or staining artifacts.
An alternative approach leverages cells' transparency itself. When light passes through transparent cells, it refracts. By recording this refracted light, scientists create holograms that require reconstruction using specialized algorithms. This is where challenges emerge.
The dose makes the poison - even with light. This phototoxicity phenomenon means light exposure affects cell state and behaviour, potentially distorting the true cellular reality. WUT researchers therefore develop techniques achieving satisfactory imaging quality using minimal photons.
Every measurement - even with precise optical systems - faces noise interference. With few photons, noise becomes particularly problematic: fewer photons mean greater noise contribution. Using indivisible photons reveals the discrete, quantum nature of this challenge, characterized by shot noise and quantization noise. Additionally, holography inherently creates twin image artifacts that degrade final image quality. These were key challenges successfully addressed by QCI Lab.
IGA Algorithm Changes the Game
Our team developed the iterative Gabor averaging (IGA) algorithm - an innovative approach combining iterative phase retrieval with frame averaging to effectively suppress both twin image artifacts and shot noise in multi-frame digital in-line holographic microscopy (DIHM).
IGA works through an iterative process reconstructing high-fidelity phase images while selectively averaging detector shot noise between consecutive frames. Simulations show IGA outperforms conventional methods, achieving better reconstruction accuracy especially under high-noise conditions. Experimental studies - including rapid imaging of motile sperm and measurements of static phase test objects under low illumination - confirmed IGA's effectiveness. The algorithm also works well for optically thin samples that produce low signal-to-noise holograms even with high photon budgets. These achievements make IGA a powerful tool for low-photostimulation rapid imaging of dynamic biological samples and for imaging extremely thin samples, potentially revolutionizing biomedical and environmental applications under low-light conditions.
"Our algorithm enables microscopic imaging with even lower illumination intensity than before," emphasizes Mikołaj Rogalski, PhD, the publication's lead author. "This proves crucial when imaging living cell colonies, as even low illumination can stimulate or damage cells, distorting measurement results."
This hybrid approach delivers higher-quality phase reconstruction even at very low illumination. IGA minimizes phototoxicity and enables rapid imaging of dynamic biological samples. Our scientists maintain momentum and already focus on new challenges.
"Our next goals include further noise reduction under low-intensity imaging conditions," reveals Rogalski, PhD. "We also plan algorithm applications in actual biological studies to minimize phototoxicity's impact on cell migration analysis."
The Quantitative Computational Imaging Lab (QCI Lab) specialises in advanced imaging techniques, combining modern numerical reconstruction algorithms with experimental optical systems. The team focuses on high-resolution optical microscopy - including interferometric techniques, lensless holographic microscopy, Fourier ptychography, and differential phase contrast - with emphasis on quantitative, non-invasive imaging. Research findings already see broad applications in biology and medicine. The group is led by Prof. Maciej Trusiak and Piotr Zdańkowski, PhD, with coordination by Błażej Roch Żyliński.