Biomedical Image Processing with Containers and Deep Learning: An Automated Analysis Pipeline

Abstract

Here, a streamlined, scalable, laboratory approach is discussed that enables medium‐to‐large dataset analysis. The presented approach combines data management, artificial intelligence, containerization, cluster orchestration, and quality control in a unified analytic pipeline. The unique combination of these individual building blocks creates a new and powerful analysis approach that can readily be applied to medium‐to‐large datasets by researchers to accelerate the pace of research. The proposed framework is applied to a project that counts the number of plasmonic nanoparticles bound to peripheral blood mononuclear cells in dark‐field microscopy images. By using the techniques presented in this article, the images are automatically processed overnight, without user interaction, streamlining the path from experiment to conclusions.

Characterization of Cell-Bound CA125 on Immune Cell Subtypes of Ovarian Cancer Patients Using a Novel Imaging Platform

“…Such split was done once and never changed. We trained the model using the adaptive momentum optimization algorithm optimizing the Dice [ 22 ] coefficient, analogously to [ 23 ]. Training was performed for 150 epochs.…”

“…PNP quantification on cells followed the methodology of [ 23 ]. Briefly, for each cell, we extract a centered 3D region of interest.…”

Translational biophotonics with Raman imaging: clinical applications and beyond

“…These automated data analysis approaches that leverage machine learning will likely be extremely important in sifting through such high dimensionality clinical datasets. 64,102 Efforts thus far are very much the tip of the iceberg: major advances in Raman data analysis techniques have begun to focus on handling the volumes of high-dimensional data generated during a single experiment, yet the expected developments that continue to focus on these efforts in the near-term will drive future clinical translation.…”

Density Distribution Maps: A Novel Tool for Subcellular Distribution Analysis and Quantitative Biomedical Imaging

“…On the other hand, image-based techniques with proper algorithms can often compensate for insufficient instrumentations. In fact, image processing has become an integral tool in the daily activity of most laboratories, where it mainly serves the automation of procedures that have been manual for many years, thus providing fast, quantitative and repeatable measurements of imaged structures descriptors, such as object's dimension and shape [26]. As a main example, preclinical in vitro studies, which typically assay drug efficacy and effectiveness by mean of commercial kits or user-validated protocols, have been profitably integrated with microscopy imaging as a further tool to complete the preclinical evaluation of antitumoral effects of the treatment investigated [27,28].…”