Super Resolution Imaging in 2D and 3D Using the SUPPOSe Algorithm Based on Approximating the Imaged Object by a Superposition of Point Sources

Abstract

A new method has been developed that opens the field of super-resolution microscopy to standard equipment and single shot experiments. In this manner fast events can be traced with high spatial resolution never available before. This new method, SUPPOSe (for Superposition of point sources) is based on the assumption that the target object to be retrieved (hidden by the resolution of the instrument) can be represented as the superposition of virtual sources of the same intensity. In this manner the ill posed mathematical problem of inverting the convolution (deconvolution) is converted to a well posed problem and a resolution beyond the instrument limitation can be achieved. The limits to this deconvolution are placed by the noise figure of the measurement and the precision with which the instrument response function is determined. The method is thoroughly tested using synthetic images and is validated with the super-resolution of intracellular structures such as mithocondria, actin filaments and microtubules. An extension to scattering images (non fluorescent) is presented and several improvements in the algorithm are described including the use of neural networks for image denoising before processing with SUPPOSe.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 09, 1921
Accession Number
AD1154647

Entities

People

  • Oscar Martinez

Tags

Communities of Interest

  • Air Platforms
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Acquisition
  • Artificial Intelligence
  • Cells
  • Change Detection
  • Compressed Sensing
  • Construction
  • Diffraction
  • Endothelial Cells
  • Frequency
  • Genetic Algorithms
  • Microscopes
  • Microscopy
  • Neural Networks
  • Numerical Aperture
  • Parallel Computing
  • Processing Equipment
  • Scattering
  • Signal Processing
  • Three Dimensional

Readers

  • Computer Vision.
  • Operations Research
  • Wave Propagation and Nonlinear Chaotic Dynamics.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks