Autonomous adaptive data acquisition for scanning hyperspectral imaging

Area of Science:

• Biophysics
• Chemical Imaging
• Microscopy

Background:

• Spectral microscopy offers valuable biochemical insights but suffers from long acquisition times, limiting the study of dynamic biological processes.
• Current methods generate high-dimensional data, requiring extensive time for image acquisition, hindering real-time analysis.
• Observing transient biological events is challenging due to the slow speed of traditional spectromicroscopy.

Purpose of the Study:

• To develop an efficient, autonomous, grid-less adaptive sampling method for spectromicroscopy.
• To accelerate image acquisition while enhancing data density in critical regions.
• To enable real-time spatiochemical imaging for biological applications.

Main Methods:

• Implemented a grid-less autonomous adaptive sampling strategy with scanning Fourier Transform infrared spectromicroscopy.
• Compared the adaptive sampling method against standard uniform grid sampling.
• Utilized performance metrics and multivariate infrared spectral analysis for quantitative and qualitative assessment.

Main Results:

• The grid-less adaptive sampling method significantly decreased image acquisition time.
• Increased sampling density in areas with steep physico-chemical gradients was achieved.
• The method demonstrated superior performance in both a two-component chemical model and a complex biological sample (Caenorhabditis elegans).

Conclusions:

• The developed adaptive sampling method enhances the efficiency of spectromicroscopy for spatiochemical imaging.
• This approach overcomes the limitations of long acquisition times, paving the way for real-time biological process observation.
• The technique shows promise for advancing biochemical analysis in complex biological systems.