Accurate and Robust Trypan Blue-Based Cell Viability Measurement Using Neural Networks

Published

July 19, 2021

Author(s)

Adele Peskin, Steven Lund, Chenyi Ling, Laura Pierce, Sumona Sarkar, Firdavs Kurbanov, Michael Halter, Joe Chalfoun, John T. Elliott

Abstract

Trypan blue dye exclusion-based cell viability measurements are highly dependent upon image quality and consistency. In order to make measurements repeatable, one must be able to reliably capture images at a consistent focal plane, and with signal-to-noise ratio within appropriate limits to support proper execution of image analysis routines. Using neural networks to determine whether cells in a bright field image are dead or alive has done away with the need to return to the sharpest image with each new sample in order to take a high-quality measurement. The algorithm is no longer trying to find human-specified image features. Instead, neural networks use convolutional layers to detect the relevant feature information to detect and classify single cells as live or dead. Our viability measurements can be made over a wide range of focal planes (up to 150 µm), and viability levels (0-80 % viability in test sets), while keeping the viability estimates to within that made using manual labeling.

Proceedings Title

CVPR 2021 open access

Conference Dates

June 19-25, 2021

Conference Location

Virtual, CO, US

Conference Title

Computer Vision for Microscopy Image Analysis (CVMI)

Pub Type

Conferences

Download Paper

Local Download

Keywords

cell viability, neural networks

Citation

Peskin, A. , Lund, S. , Ling, C. , Pierce, L. , Sarkar, S. , Kurbanov, F. , Halter, M. , Chalfoun, J. and Elliott, J. (2021), Accurate and Robust Trypan Blue-Based Cell Viability Measurement Using Neural Networks, CVPR 2021 open access, Virtual, CO, US, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=932024 (Accessed April 28, 2024)