Machine Learning Approach to Estimate Volumetric Quantification of 2D SEM Images of Biofilms

Scanning Electron Microscopy (SEM) is one of the most important imaging techniques to understand the dynamics of microscale objects. However, SEM images could mostly produce two-dimensional images which limit the exploration of volumetric information. To address this limitation, we propose the integration of a state-of-the-art machine learning approach that leverages Confocal Laser Scanning Microscopy (CLSM) image data for training a deep learning model. The objective of this model is to approximate volumetric information by taking a 2D SEM image as input and generating corresponding depth information. The depth of information then be utilized to estimate volumetric quantification such as biofilm density and bacteria cell counts.