Loading...
 
Skip to main content
Scientific Volume Imaging B.V.

Scientific Volume Imaging

Deconvolution - Visualization - Analysis


Huygens' Hybrid AI filament deconvolution

Huygens' Hybrid AI deconvolution approach for filaments combines traditional image restoration techniques with advanced artificial intelligence to significantly enhance the clarity and resolution of filamentous structures in microscopy images. This method leverages the robustness of classic deconvolution algorithms with the pattern recognition and adaptability of AI, enabling more precise reconstruction of fine, thread-like features such as endoplasmic reticulum or cytoskeletal filaments. By doing so, it minimizes noise and artifacts while preserving structural integrity, making it a powerful tool for researchers seeking high-fidelity visualization and quantification of delicate biological structures. Curious about Huygens' other AI solutions? See here!

Hybrid AI Deconvolution
Raw data
Hybrid AI Filament Deconvolution Huygens' Hybrid AI deconvolution optimally resolves filamentous structures. Here, a raw low signal image of microtubules (left) was deconvolved using Huygens Hybrid AI (right). Courtesy of Robin Pelle and Prof. Lukas Kapitein, Cell biology, Utrecht University, The Netherlands. Image was captured with a Zeiss Airyscan® microscope.
Hybrid AI Deconvolution
Classic Deconvolution

How does it compare to Classic Deconvolution?


The hybrid AI deconvolution works by including a priori information in the iterative maximum likelyhood deconvolution process, resulting in a bias of neighboring point spread functions on filaments towards a more even filament structure.
The Huygens Hybrid AI Deconvolution can be accessed in the Deconvolution Express by clicking "Adjust acuity" and selecting "Filaments" as strategy.

Hybrid AI Deconvolution versus Classic Deconvolution: Hybrid AI deconvolution (right) performs better for filamentous structures, such as this tubulin staining, compared to classic deconvolution (left) in Huygens. The non-filamentous structures such as the nucleus are not affected. Courtesy of Robin Pelle and Prof. Lukas Kapitein, Cell biology, Utrecht University, The Netherlands. Captured by confocal microscope.

Hybrid AI deconvolution on low signal images.


Huygens Hybrid AI deconvolution allows sampling with lower acquisition settings and still get comparable results to higher sampled images. This prevents common issues such as phototoxicity and bleaching.

Hybrid AI Filament Deconvolution Huygens' Hybrid AI deconvolution allows you to optimally visualize filamentous structures. Here images stained for tubulin were deconvolved using Hybrid AI (right), compared to the raw data (left). Comparing the low signal (top) with the high signal (bottom) images shows that the Hybrid AI deconvolution on the low signal image gives comparable results to the high signal image. Courtesy of Robin Pelle and Prof. Lukas Kapitein, Cell biology, Utrecht University, The Netherlands. Captured by confocal microscope.


Signal Comparison