Label-Free Quantification of Gold Nanoparticles at the Single-Cell level using MC-CNN

Paper Links

Abstract

Gold Nanoparticles (AuNPs) are extensively used in cellular imaging, single particle tracking, disease detection, membrane protein interaction and drug delivery. However, it is difficult to analyze the interactions and uptake of AuNPs in live cells. Optical or electron microscopy-based imaging techniques are destructive, and moreover the resultant images can feature hundreds of particles with highly varying intensity levels, blurring and major occlusion that are difficult to count visually. To address this issue and quantify AuNP uptake within live cells from images, we annotated a dataset of dark-field images of 50 nanometer radius AuNPs at different incubation times in HeLa cells. Then we designed a custom multi-column convolutional neural network to count the number of particles present within a cell. We achieved better results compared to existing cell counting architectures and compared it to spectroscopy-based counting. Furthermore, we are disseminating a dataset of dot-instance annotated AuNPs to support further development of dark-field nanoparticle image processing and analysis.

Code

Available on Github

Dataset

Please contact the authors for access to the dataset. We may make it publicly available at some point in the future.

Credits

Shadab Hafiz Choudhury¹ and Dr. Abu Mohsin^{1 2}

^{1: Department of Electrical and Electronic Engineering, Brac University}
^{2: Corresponding Author}
_{Both authors contributed equally to this work.}