Nanoparticles occur in various environments as a consequence of man-made processes, which raises concerns about their impact on the environment and human health. To allow for proper risk assessment, a precise and statistically relevant analysis of particle characteristics (such as e.g. size, shape and composition) is required that would greatly benefit from automated image analysis procedures. While deep learning shows impressive results in object detection tasks, its applicability is limited by the amount of representative, experimentally collected and manually annotated training data. Here, we present an elegant, flexible and versatile method to bypass this costly and tedious data acquisition process. We show that using a rendering software allows to generate realistic, synthetic training data to train a state-of-the art deep neural network. Using this approach, we derive a segmentation accuracy that is comparable to man-made annotations for toxicologically relevant metal-oxide nanoparticle ensembles which we chose as examples. Our study paves the way towards the use of deep learning for automated, high-throughput particle detection in a variety of imaging techniques such as microscopies and spectroscopies, for a wide variety of studies and applications, including the detection of plastic micro- and nanoparticles.
@article{arxiv.2011.10505,
title = {Synthetic Image Rendering Solves Annotation Problem in Deep Learning Nanoparticle Segmentation},
author = {Leonid Mill and David Wolff and Nele Gerrits and Patrick Philipp and Lasse Kling and Florian Vollnhals and Andrew Ignatenko and Christian Jaremenko and Yixing Huang and Olivier De Castro and Jean-Nicolas Audinot and Inge Nelissen and Tom Wirtz and Andreas Maier and Silke Christiansen},
journal= {arXiv preprint arXiv:2011.10505},
year = {2020}
}