Monitoring of Wild Pseudomonas Biofilm Strain Conditions Using Statistical Characterisation of Scanning Electron Microscopy Images

TL;DR

This paper introduces a statistical image analysis method to quantify biofilm architecture variations under different growth conditions, aiming to develop automated monitoring for medical implants.

Contribution

It presents a novel approach using texture and fractal analysis of SEM images to predict biofilm growth conditions on polymer surfaces with plasma protein conditioning.

Findings

Identified discriminatory features for different biofilm growth conditions.

Demonstrated potential for automated biofilm monitoring in medical settings.

Analyzed effects of substrate and conditioning layer variations.

Abstract

The present paper proposes a novel method of quantification of the variation in biofilm architecture, in correlation with the alteration of growth conditions that include, variations of substrate and conditioning layer. The polymeric biomaterial serving as substrates are widely used in implants and indwelling medical devices, while the plasma proteins serve as the conditioning layer. The present method uses descriptive statistics of FESEM images of biofilms obtained during a variety of growth conditions. We aim to explore here the texture and fractal analysis techniques, to identify the most discriminatory features which are capable of predicting the difference in biofilm growth conditions. We initially extract some statistical features of biofilm images on bare polymer surfaces, followed by those on the same substrates adsorbed with two different types of plasma proteins, viz. Bovine serum albumin (BSA) and Fibronectin (FN), for two different adsorption times. The present analysis has the potential to act as a futuristic technology for developing a computerized monitoring system in hospitals with automated image analysis and feature extraction, which may be used to predict the growth profile of an emerging biofilm on surgical implants or similar medical applications.