An image-processing method to detect sub-optical features based on understanding noise in intensity measurements

Tripta Bhatia

doi:10.1007/s00249-017-1273-z

Abstract

Accurate quantitative analysis of image data requires that we distinguish between fluorescence intensity (true signal) and the noise inherent to its measurements to the extent possible. We image multilamellar membrane tubes and beads that grow from defects in the fluid lamellar phase of the lipid 1,2-dioleoyl-sn-glycero-3-phosphocholine dissolved in water and water-glycerol mixtures by using fluorescence confocal polarizing microscope. We quantify image noise and determine the noise statistics. Understanding the nature of image noise also helps in optimizing image processing to detect sub-optical features, which would otherwise remain hidden. We use an image-processing technique “optimum smoothening” to improve the signal-to-noise ratio of features of interest without smearing their structural details. A high SNR renders desired positional accuracy with which it is possible to resolve features of interest with width below optical resolution. Using optimum smoothening, the smallest and the largest core diameter detected is of width 88 pm 23 and 6860 pm 50 nm, respectively, discussed in this paper. The image-processing and analysis techniques and the noise modeling discussed in this paper can be used for detailed morphological analysis of features down to sub-optical length scales that are obtained by any kind of fluorescence intensity imaging in the raster mode.

Highlights

Membrane tubes are found in cells connecting different membrane compartments (Lee and Chen 1988)
We restrict our discussion to 1-D variation of the fluorescence intensity along y at fixed x
If [ Iav(y) ] is decreased together with [ (y) ] this eventually results in a decrease in signal-to-noise ratio (SNR)

Summary

Introduction

Membrane tubes are found in cells connecting different membrane compartments (Lee and Chen 1988). We have discussed membrane tubes (Bhatia et al 2015) of the lipid, 1,2-dioleoyl-sn-glycero3-phosphocholine (DOPC) that grow from the defects in the multilamellar lipid reservoir hydrated in excess water. We have found that near the neck of the beads, the inner core of the tubes is not blocked by lipid material, i.e., the inner core runs continuously through the tubes and the beads.

Methods

Results

Conclusions