infocus #65 March 2022 2021 Summer Studentship Report: Image processing and analysis methods for the detection of islets of Mesolens images of whole mouse pancreas specimens

DOI: 10.22443/rms.inf.1.220

This project was carried out remotely and was entirely computational, using image datasets produced by Ms Katrina Wesencraft at the University of Strathclyde.

The majority of the work was undertaken using FIJI [1], an image processing and analysis programme, using various plugins. One of the key plugins used, CLAHE [2], is a tool that breaks the image into multiple small sections, and increases the contrast within each, resulting in a locally contrasted image as opposed to carrying out the procedure on the whole image, and therefore overly contrast-adjusting some regions.

Alternative plugins such as Trackmate [3], a plugin designed for particle detection and tracking, had some success for islet detection as this sought out intensely fluorescent areas of a circular shape.

The in-built FIJI functions could also be used for islet detection, as once an intensity threshold was applied to the data, it was possible to use the “Analyse particles” function to detect circular shapes within images, based upon an approximate area.