Presentation
7 March 2022 Nuclei detection in whole embryonic hearts with a fully convolutional regression network
Author Affiliations +
Abstract
Advances in tissue clearing and three-dimensional microscopy require new tools to analyze the resulting large volumes with single-cell resolution. Many existing nuclei detection approaches fail when applied to the developing heart, with its high cell density, and elongated myocytes. We propose a new regression-based convolutional neural network that detect nuclei centroids in whole DAPI-stained embryonic quail hearts. High nuclei detection accuracy was obtained in two different hearts where our algorithm outperformed other deep learning approaches. Once nuclei were identified we were also able to extract properties such as orientation and size, which enables future studies of heart development and disease.
Conference Presentation
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Maryse Lapierre-Landry, Zexuan Liu, Shan Ling, Mahdi Bayat, David L. Wilson, and Michael W. Jenkins "Nuclei detection in whole embryonic hearts with a fully convolutional regression network", Proc. SPIE PC11936, Diagnostic and Therapeutic Applications of Light in Cardiology 2022, PC119360A (7 March 2022); https://doi.org/10.1117/12.2608785
Advertisement
Advertisement
KEYWORDS
Heart

3D image processing

Image analysis

Tissues

Image processing algorithms and systems

Image resolution

Image segmentation

Back to Top